arabic_plate.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Collecting imutils\n",
      "  Downloading imutils-0.5.4.tar.gz (17 kB)\n",
      "  Preparing metadata (setup.py): started\n",
      "  Preparing metadata (setup.py): finished with status 'done'\n",
      "Building wheels for collected packages: imutils\n",
      "  Building wheel for imutils (setup.py): started\n",
      "  Building wheel for imutils (setup.py): finished with status 'done'\n",
      "  Created wheel for imutils: filename=imutils-0.5.4-py3-none-any.whl size=25855 sha256=43685ee228cbb0fac9c6d0d85833eb971ec50e2cf4633e63efeeae050672c1b4\n",
      "  Stored in directory: c:\\users\\aravi\\appdata\\local\\pip\\cache\\wheels\\85\\cf\\3a\\e265e975a1e7c7e54eb3692d6aa4e2e7d6a3945d29da46f2d7\n",
      "Successfully built imutils\n",
      "Installing collected packages: imutils\n",
      "Successfully installed imutils-0.5.4\n",
      "Note: you may need to restart the kernel to use updated packages.\n"
     ]
    }
   ],
   "source": [
    "pip install imutils"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from ultralytics import YOLO\n",
    "import tensorflow as tf\n",
    "import cv2\n",
    "import os\n",
    "import imutils\n",
    "import cv2\n",
    "import itertools, os, time\n",
    "import numpy as np\n",
    "import argparse\n",
    "import pandas as pd\n",
    "import operator\n",
    "import os, shutil\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Build model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0123456789ABCDEFGHIJKLMNOPQRSTUVWYZabnX\n"
     ]
    }
   ],
   "source": [
    "import string\n",
    "\n",
    "# Original mapping from class_name to English letters\n",
    "original_mapping = {\n",
    "    \"1\": \"1\", \"6\": \"6\", \"5\": \"5\", \"4\": \"4\", \"7\": \"7\", \"9\": \"9\", \"3\": \"3\", \"2\": \"2\", \"8\": \"8\",\n",
    "    \"meem\": \"م\", \"0\": \"0\", \"raa\": \"ر\", \"seen\": \"س\", \"alif\": \"ا\", \"daal\": \"د\", \"noon\": \"ن\",\n",
    "    \"waw\": \"و\", \"ain\": \"ع\", \"haa\": \"ه\", \"laam\": \"ل\", \"jeem\": \"ج\", \"baa\": \"ب\", \"qaaf\": \"ق\",\n",
    "    \"yaa\": \"ى\", \"faa\": \"ف\", \"Taa\": \"ط\", \"saad\": \"ص\", \"thaa\": \"ث\", \"ghayn\": \"ج\", \"sheen\": \"ش\",\n",
    "    \"Thaa\": \"ظ\", \"khaa\": \"خ\", \"kaaf\": \"ك\", \"taa\": \"ت\", \"zay\": \"ز\", \"zaal\": \"ذ\", \"7aa\": \"ح\",\n",
    "    \"daad\": \"ض\"\n",
    "}\n",
    "\n",
    "# Arabic to English mapping\n",
    "ar_to_en = {\n",
    "    \"١\": \"1\", \"٢\": \"2\", \"٣\": \"3\", \"٤\": \"4\", \"٥\": \"5\", \"٦\": \"6\", \"٧\": \"7\", \"٨\": \"8\", \"٩\": \"9\", \"٠\": \"0\",\n",
    "    \"ا\": \"A\", \"ب\": \"B\", \"ت\": \"C\", \"ث\": \"D\", \"ج\": \"E\", \"ح\": \"F\", \"خ\": \"G\", \"د\": \"H\", \"ذ\": \"I\",\n",
    "    \"ر\": \"J\", \"ز\": \"K\", \"س\": \"L\", \"ش\": \"M\", \"ص\": \"N\", \"ض\": \"O\", \"ط\": \"P\", \"ظ\": \"Q\", \"ع\": \"R\", \"غ\": \"S\",\n",
    "    \"ف\": \"T\", \"ق\": \"U\", \"ك\": \"V\", \"ل\": \"W\", \"م\": \"n\", \"ن\": \"Y\", \"ه\": \"Z\", \"و\": \"a\", \"ى\": \"b\",\n",
    "    \"1\": \"1\", \"2\": \"2\", \"3\": \"3\", \"4\": \"4\", \"5\": \"5\", \"6\": \"6\", \"7\": \"7\", \"8\": \"8\", \"9\": \"9\", \"0\": \"0\"\n",
    "}\n",
    "\n",
    "en_to_ar = {v: k for k, v in ar_to_en.items()}\n",
    "\n",
    "# Remove duplicate values to get unique letters\n",
    "unique_letters = sorted(set(ar_to_en.values()))\n",
    "\n",
    "# Add 'X' as the padding character\n",
    "unique_letters.append('X')\n",
    "\n",
    "# Convert the unique letters list to a string\n",
    "CHAR_VECTOR = ''.join(unique_letters)\n",
    "\n",
    "# Create a list of letters\n",
    "letters = [letter for letter in CHAR_VECTOR]\n",
    "\n",
    "num_classes = len(letters) + 1\n",
    "\n",
    "img_w, img_h = 128, 64\n",
    "\n",
    "# Network parameters\n",
    "batch_size = 8\n",
    "val_batch_size = 1\n",
    "\n",
    "downsample_factor = 4\n",
    "max_text_len = 7\n",
    "\n",
    "print(CHAR_VECTOR)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Model Layers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import tensorflow.keras as keras\n",
    "from tensorflow.keras.layers import Conv2D, MaxPooling2D\n",
    "from tensorflow.keras.layers import Dense, Dropout, Activation, Flatten, Input\n",
    "from tensorflow.keras.layers import BatchNormalization, Reshape, Lambda\n",
    "from tensorflow.keras.layers import LSTM, Bidirectional\n",
    "from tensorflow.keras.layers import add, concatenate\n",
    "\n",
    "\n",
    "# VGG Blocks\n",
    "class VggBlock1(keras.layers.Layer):\n",
    "    def __init__(self, num_filters, do_maxpool: bool = True, **kwargs):\n",
    "        super(VggBlock1, self).__init__(**kwargs)\n",
    "        self.num_filters = num_filters\n",
    "        self.do_maxpool = do_maxpool\n",
    "        self.conv = Conv2D(\n",
    "            num_filters, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\"\n",
    "        )\n",
    "        self.bn = BatchNormalization()\n",
    "        self.relu = Activation(\"relu\")\n",
    "        if self.do_maxpool:\n",
    "            self.pool = MaxPooling2D(pool_size=(2, 2))\n",
    "\n",
    "    def call(self, inputs):\n",
    "        x = self.conv(inputs)\n",
    "        x = self.bn(x)\n",
    "        x = self.relu(x)\n",
    "        if self.do_maxpool:\n",
    "            x = self.pool(x)\n",
    "        return x\n",
    "\n",
    "    def get_config(self):\n",
    "        config = super(VggBlock1, self).get_config()\n",
    "        config.update({\n",
    "            'num_filters': self.num_filters,\n",
    "            'do_maxpool': self.do_maxpool\n",
    "        })\n",
    "        return config\n",
    "\n",
    "\n",
    "class VggBlock2(keras.layers.Layer):\n",
    "    def __init__(self, num_filters, **kwargs):\n",
    "        super(VggBlock2, self).__init__(**kwargs)\n",
    "        self.num_filters = num_filters\n",
    "        \n",
    "        self.conv1 = Conv2D(\n",
    "            num_filters, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\"\n",
    "        )\n",
    "        self.bn1 = BatchNormalization()\n",
    "        self.relu1 = Activation(\"relu\")\n",
    "        self.conv2 = Conv2D(\n",
    "            num_filters, (3, 3), padding=\"same\", kernel_initializer=\"he_normal\"\n",
    "        )\n",
    "        self.bn2 = BatchNormalization()\n",
    "        self.relu2 = Activation(\"relu\")\n",
    "        self.pool = MaxPooling2D(pool_size=(1, 2))\n",
    "\n",
    "    def call(self, inputs):\n",
    "        x = self.conv1(inputs)\n",
    "        x = self.bn1(x)\n",
    "        x = self.relu1(x)\n",
    "        x = self.conv2(x)\n",
    "        x = self.bn2(x)\n",
    "        x = self.relu2(x)\n",
    "        x = self.pool(x)\n",
    "        return x\n",
    "    \n",
    "    def get_config(self):\n",
    "        config = super(VggBlock2, self).get_config()\n",
    "        config.update({\"num_filters\": self.num_filters})\n",
    "        return config\n",
    "\n",
    "\n",
    "\n",
    "# RNN\n",
    "class BI_LSTM_Block(keras.layers.Layer):\n",
    "    def __init__(self, num_units, merge_mode, **kwargs):\n",
    "        super(BI_LSTM_Block, self).__init__(**kwargs)\n",
    "        self.num_units = num_units\n",
    "        self.merge_mode = merge_mode\n",
    "\n",
    "        self.bi_lstm = Bidirectional(\n",
    "            LSTM(num_units, return_sequences=True, kernel_initializer=\"he_normal\"),\n",
    "            merge_mode=merge_mode,\n",
    "        )\n",
    "        self.bn = BatchNormalization()\n",
    "\n",
    "    def call(self, inputs):\n",
    "        x = self.bi_lstm(inputs)\n",
    "        x = self.bn(x)\n",
    "        return x\n",
    "\n",
    "    def get_config(self):\n",
    "        config = super(BI_LSTM_Block, self).get_config()\n",
    "        config.update({\"num_units\": self.num_units, \"merge_mode\": self.merge_mode})\n",
    "        return config\n",
    "\n",
    "\n",
    "\n",
    "def ctc_lamda_func(args):\n",
    "    y_pred, labels, input_length, label_length = args\n",
    "\n",
    "    y_pred = y_pred[:, 2:, :]\n",
    "    return keras.backend.ctc_batch_cost(labels, y_pred, input_length, label_length)\n",
    "\n",
    "\n",
    "def build_model(training):\n",
    "\n",
    "    input_shape = (128, 64, 1)\n",
    "\n",
    "    inputs = Input(name=\"the_input\", shape=input_shape, dtype=\"float32\")\n",
    "\n",
    "    # VGG (CNN)\n",
    "    x = VggBlock1(64, name=\"VGG_Block1\")(inputs)\n",
    "    x = VggBlock1(128, name=\"VGG_Block2\")(x)\n",
    "    x = VggBlock2(256, name=\"VGG_Block3\")(x)\n",
    "    x = VggBlock2(512, name=\"VGG_Block4\")(x)\n",
    "    x = VggBlock1(512, do_maxpool=False, name=\"VGG_Block5\")(x)\n",
    "\n",
    "    # CNN to RNN\n",
    "    new_shape = (32, 2048)\n",
    "    x = Reshape(target_shape=new_shape, name=\"reshape\")(x)\n",
    "    x = Dense(64, activation=\"relu\", kernel_initializer=\"he_normal\", name=\"dense1\")(x)\n",
    "\n",
    "    # RNN\n",
    "    x = BI_LSTM_Block(256, merge_mode=\"sum\", name=\"BI_LSTM_Block1\")(x)\n",
    "    x = BI_LSTM_Block(128, merge_mode=\"concat\", name=\"BI_LSTM_Block2\")(x)\n",
    "\n",
    "    x = Dropout(0.25)(x)\n",
    "\n",
    "    # RNN output to character activations\n",
    "    x = Dense(num_classes, kernel_initializer=\"he_normal\", name=\"dense2\")(x)\n",
    "    y_pred = Activation(\"softmax\", name=\"softmax\")(x)\n",
    "\n",
    "    labels = Input(name=\"the_labels\", shape=[max_text_len], dtype=\"float32\")\n",
    "    input_length = Input(name=\"input_length\", shape=[1], dtype=\"int64\")\n",
    "    label_length = Input(name=\"label_length\", shape=[1], dtype=\"int64\")\n",
    "\n",
    "    loss_out = Lambda(ctc_lamda_func, output_shape=(1,), name=\"ctc\")(\n",
    "        [y_pred, labels, input_length, label_length]\n",
    "    )\n",
    "\n",
    "    if training:\n",
    "        return keras.Model(\n",
    "            inputs=[inputs, labels, input_length, label_length], outputs=loss_out\n",
    "        )\n",
    "    else:\n",
    "        return keras.Model(inputs=[inputs], outputs=y_pred)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Model Checkpoints"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint\n",
    "\n",
    "# Define the EarlyStopping and ModelCheckpoint callbacks\n",
    "early_stopping = EarlyStopping(monitor='val_loss', patience=25, verbose=1, mode='min', restore_best_weights=True)\n",
    "model_checkpoint = ModelCheckpoint('model_ad_best.keras', monitor='val_loss', mode='min', save_best_only=True, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From c:\\Users\\aravi\\anaconda3\\envs\\v8OPS\\lib\\site-packages\\keras\\src\\backend\\tensorflow\\core.py:204: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "model = build_model(True)\n",
    "opt = tf.keras.optimizers.Adadelta(1.0)\n",
    "model.compile(loss={'ctc': lambda y_true, y_pred: y_pred}, metrics=['accuracy'], optimizer=opt)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"functional\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"functional\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)        </span>┃<span style=\"font-weight: bold\"> Output Shape      </span>┃<span style=\"font-weight: bold\">    Param # </span>┃<span style=\"font-weight: bold\"> Connected to      </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━┩\n",
       "│ the_input           │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>,   │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                 │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │ <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)                │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ VGG_Block1          │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>,    │        <span style=\"color: #00af00; text-decoration-color: #00af00\">896</span> │ the_input[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]   │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">VggBlock1</span>)         │ <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)               │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ VGG_Block2          │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>,    │     <span style=\"color: #00af00; text-decoration-color: #00af00\">74,368</span> │ VGG_Block1[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]  │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">VggBlock1</span>)         │ <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>)              │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ VGG_Block3          │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">8</span>,     │    <span style=\"color: #00af00; text-decoration-color: #00af00\">887,296</span> │ VGG_Block2[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]  │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">VggBlock2</span>)         │ <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)              │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ VGG_Block4          │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">4</span>,     │  <span style=\"color: #00af00; text-decoration-color: #00af00\">3,544,064</span> │ VGG_Block3[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]  │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">VggBlock2</span>)         │ <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)              │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ VGG_Block5          │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">4</span>,     │  <span style=\"color: #00af00; text-decoration-color: #00af00\">2,361,856</span> │ VGG_Block4[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]  │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">VggBlock1</span>)         │ <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)              │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ reshape (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Reshape</span>)   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">2048</span>)  │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ VGG_Block5[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]  │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)    │    <span style=\"color: #00af00; text-decoration-color: #00af00\">131,136</span> │ reshape[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]     │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ BI_LSTM_Block1      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)   │    <span style=\"color: #00af00; text-decoration-color: #00af00\">658,432</span> │ dense1[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]      │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BI_LSTM_Block</span>)     │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ BI_LSTM_Block2      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)   │    <span style=\"color: #00af00; text-decoration-color: #00af00\">395,264</span> │ BI_LSTM_Block1[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>… │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BI_LSTM_Block</span>)     │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dropout (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)   │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ BI_LSTM_Block2[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>… │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">40</span>)    │     <span style=\"color: #00af00; text-decoration-color: #00af00\">10,280</span> │ dropout[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]     │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ softmax             │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">40</span>)    │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ dense2[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>]      │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Activation</span>)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ the_labels          │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">7</span>)         │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                 │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ input_length        │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)         │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                 │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ label_length        │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)         │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ -                 │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ ctc (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Lambda</span>)        │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)         │          <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │ softmax[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>],    │\n",
       "│                     │                   │            │ the_labels[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>], │\n",
       "│                     │                   │            │ input_length[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">…</span> │\n",
       "│                     │                   │            │ label_length[<span style=\"color: #00af00; text-decoration-color: #00af00\">0</span>][<span style=\"color: #00af00; text-decoration-color: #00af00\">…</span> │\n",
       "└─────────────────────┴───────────────────┴────────────┴───────────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)       \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape     \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m   Param #\u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mConnected to     \u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━┩\n",
       "│ the_input           │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m128\u001b[0m, \u001b[38;5;34m64\u001b[0m,   │          \u001b[38;5;34m0\u001b[0m │ -                 │\n",
       "│ (\u001b[38;5;33mInputLayer\u001b[0m)        │ \u001b[38;5;34m1\u001b[0m)                │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ VGG_Block1          │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m64\u001b[0m, \u001b[38;5;34m32\u001b[0m,    │        \u001b[38;5;34m896\u001b[0m │ the_input[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]   │\n",
       "│ (\u001b[38;5;33mVggBlock1\u001b[0m)         │ \u001b[38;5;34m64\u001b[0m)               │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ VGG_Block2          │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m16\u001b[0m,    │     \u001b[38;5;34m74,368\u001b[0m │ VGG_Block1[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]  │\n",
       "│ (\u001b[38;5;33mVggBlock1\u001b[0m)         │ \u001b[38;5;34m128\u001b[0m)              │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ VGG_Block3          │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m8\u001b[0m,     │    \u001b[38;5;34m887,296\u001b[0m │ VGG_Block2[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]  │\n",
       "│ (\u001b[38;5;33mVggBlock2\u001b[0m)         │ \u001b[38;5;34m256\u001b[0m)              │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ VGG_Block4          │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m4\u001b[0m,     │  \u001b[38;5;34m3,544,064\u001b[0m │ VGG_Block3[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]  │\n",
       "│ (\u001b[38;5;33mVggBlock2\u001b[0m)         │ \u001b[38;5;34m512\u001b[0m)              │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ VGG_Block5          │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m4\u001b[0m,     │  \u001b[38;5;34m2,361,856\u001b[0m │ VGG_Block4[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]  │\n",
       "│ (\u001b[38;5;33mVggBlock1\u001b[0m)         │ \u001b[38;5;34m512\u001b[0m)              │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ reshape (\u001b[38;5;33mReshape\u001b[0m)   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m2048\u001b[0m)  │          \u001b[38;5;34m0\u001b[0m │ VGG_Block5[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]  │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense1 (\u001b[38;5;33mDense\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m64\u001b[0m)    │    \u001b[38;5;34m131,136\u001b[0m │ reshape[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]     │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ BI_LSTM_Block1      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m256\u001b[0m)   │    \u001b[38;5;34m658,432\u001b[0m │ dense1[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]      │\n",
       "│ (\u001b[38;5;33mBI_LSTM_Block\u001b[0m)     │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ BI_LSTM_Block2      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m256\u001b[0m)   │    \u001b[38;5;34m395,264\u001b[0m │ BI_LSTM_Block1[\u001b[38;5;34m0\u001b[0m… │\n",
       "│ (\u001b[38;5;33mBI_LSTM_Block\u001b[0m)     │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dropout (\u001b[38;5;33mDropout\u001b[0m)   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m256\u001b[0m)   │          \u001b[38;5;34m0\u001b[0m │ BI_LSTM_Block2[\u001b[38;5;34m0\u001b[0m… │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ dense2 (\u001b[38;5;33mDense\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m40\u001b[0m)    │     \u001b[38;5;34m10,280\u001b[0m │ dropout[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]     │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ softmax             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m40\u001b[0m)    │          \u001b[38;5;34m0\u001b[0m │ dense2[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m]      │\n",
       "│ (\u001b[38;5;33mActivation\u001b[0m)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ the_labels          │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m7\u001b[0m)         │          \u001b[38;5;34m0\u001b[0m │ -                 │\n",
       "│ (\u001b[38;5;33mInputLayer\u001b[0m)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ input_length        │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)         │          \u001b[38;5;34m0\u001b[0m │ -                 │\n",
       "│ (\u001b[38;5;33mInputLayer\u001b[0m)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ label_length        │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)         │          \u001b[38;5;34m0\u001b[0m │ -                 │\n",
       "│ (\u001b[38;5;33mInputLayer\u001b[0m)        │                   │            │                   │\n",
       "├─────────────────────┼───────────────────┼────────────┼───────────────────┤\n",
       "│ ctc (\u001b[38;5;33mLambda\u001b[0m)        │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)         │          \u001b[38;5;34m0\u001b[0m │ softmax[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m],    │\n",
       "│                     │                   │            │ the_labels[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m0\u001b[0m], │\n",
       "│                     │                   │            │ input_length[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m…\u001b[0m │\n",
       "│                     │                   │            │ label_length[\u001b[38;5;34m0\u001b[0m][\u001b[38;5;34m…\u001b[0m │\n",
       "└─────────────────────┴───────────────────┴────────────┴───────────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">8,063,592</span> (30.76 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m8,063,592\u001b[0m (30.76 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">8,058,088</span> (30.74 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m8,058,088\u001b[0m (30.74 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">5,504</span> (21.50 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m5,504\u001b[0m (21.50 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Encoding and Decoding function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#map the arabic op to english letter\n",
    "ar_to_plate_map = {\n",
    "    \"ا\":\"A\", \"ب\":\"B\", \"ح\":\"J\", \"د\":\"D\", \n",
    "    \"ر\":\"R\", \"س\":\"S\", \"ص\":\"X/C\", \"ط\":\"T\", \n",
    "    \"ع\":\"E\", \"ق\":\"G/K\", \"ك\":\"K\", \"ل\":\"L\", \n",
    "    \"م\":\"Z/M\", \"ن\":\"N\", \"هـ\":\"H\", \"و\":\"U/W\", \"ى\":\"V\", \"ج\":\"G\", \"ف\":\"F\", \"ی\":\"Y\"\n",
    "}\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function to convert label to English\n",
    "def label_to_en(label):\n",
    "    new = \"\"\n",
    "    for l in label:\n",
    "        if l == \" \":\n",
    "            continue\n",
    "        elif l in string.digits:\n",
    "            new += l\n",
    "        else:\n",
    "            try:\n",
    "                new = new + ar_to_plate_map[l]\n",
    "            except:\n",
    "                continue\n",
    "    max_letters = 7\n",
    "    missing = max_letters - len(new)\n",
    "    if missing > 0:\n",
    "        new = new + \"-\" * missing\n",
    "    return new\n",
    "\n",
    "# Function to reverse a string\n",
    "def reverse_str(s):\n",
    "    return s[::-1]\n",
    "\n",
    "# Function to convert label to Arabic\n",
    "def label_to_ar(label):\n",
    "    label = reverse_str(label)\n",
    "    ar_label = \"\"\n",
    "    for i in label:\n",
    "        if i == \"X\":\n",
    "            continue\n",
    "        else:\n",
    "            ar_label += en_to_ar[i] + \" \"\n",
    "    ar_label = ar_label.strip()\n",
    "    return ar_label\n",
    "\n",
    "# Function to get label from indexes\n",
    "def get_label(idxes):\n",
    "    label = \"\"\n",
    "    for idx in idxes:\n",
    "        if idx < len(letters):\n",
    "            label += letters[idx]\n",
    "    return label\n",
    "\n",
    "# Function to decode label\n",
    "def decode_label(out):\n",
    "    out_best = list(np.argmax(out[0, 2:], axis=1))\n",
    "    out_best = [k for k, g in itertools.groupby(out_best)]\n",
    "    outstr = ''\n",
    "    for i in out_best:\n",
    "        if i < len(letters):\n",
    "            outstr += letters[i]\n",
    "    return outstr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From c:\\Users\\aravi\\anaconda3\\envs\\v8OPS\\lib\\site-packages\\keras\\src\\backend\\tensorflow\\core.py:204: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "numplate_model = build_model(False)\n",
    "numplate_model.load_weights(\"models\\\\best_weight.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#multi pred test\n",
    "\n",
    "import os\n",
    "import cv2\n",
    "from ultralytics import YOLO\n",
    "import arabic_reshaper\n",
    "from bidi.algorithm import get_display\n",
    "import numpy as np\n",
    "from PIL import ImageFont, ImageDraw, Image\n",
    "\n",
    "def numberplate_detection(image_path, yolo_model, model):\n",
    "    img = cv2.imread(image_path)\n",
    "    results = yolo_model.predict(image_path)\n",
    "    bounding_boxes = []\n",
    "    \n",
    "    # if detections were made\n",
    "    if results[0].boxes is not None:\n",
    "        for box in results[0].boxes.xyxy:\n",
    "            x1, y1, x2, y2 = map(int, box[:4])  # Get bounding box coordinates\n",
    "            bounding_boxes.append((x1, y1, x2, y2))\n",
    "    \n",
    "    labels = []\n",
    "\n",
    "    for box in bounding_boxes:\n",
    "        x1, y1, x2, y2 = box\n",
    "        crop_img = img[y1:y2, x1:x2]  # Crop the image\n",
    "        #converting to bw\n",
    "        img_grey = cv2.cvtColor(crop_img, cv2.COLOR_BGR2GRAY)\n",
    "        \n",
    "        #process the img\n",
    "        img_pred = img_grey.astype(np.float32)\n",
    "        img_pred = cv2.resize(img_pred, (128, 64))\n",
    "        img_pred = (img_pred / 255.0)\n",
    "        img_pred = img_pred.T\n",
    "        img_pred = np.expand_dims(img_pred, axis=-1)\n",
    "        img_pred = np.expand_dims(img_pred, axis=0)\n",
    "\n",
    "        #predict on img\n",
    "        net_out_value = model.predict(img_pred, verbose=1)\n",
    "        pred_text = decode_label(net_out_value)\n",
    "        pred_text_ar = label_to_ar(pred_text)\n",
    "\n",
    "        labels.append(pred_text_ar)\n",
    "\n",
    "    print(labels)\n",
    "    return bounding_boxes, labels, img\n",
    "\n",
    "\n",
    "\n",
    "        \n",
    "\n",
    "def bbox_and_labels(image, bbox, labels, save_dir, filename):\n",
    "    img_pil = Image.fromarray(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))\n",
    "    draw = ImageDraw.Draw(img_pil)\n",
    "\n",
    "    font = ImageFont.truetype(font= \"arial.ttf\", size=20)\n",
    "\n",
    "    for box, label in zip(bbox,labels):\n",
    "        x1, y1, x2, y2 = box\n",
    "\n",
    "        #bbox\n",
    "        draw.rectangle([x1, y1, x2, y2], outline=(0,255,0), width=2)\n",
    "\n",
    "        #arabic text label\n",
    "        reshaped_ar_label = arabic_reshaper.reshape(label)\n",
    "        bidi_ar_text = get_display(reshaped_ar_label)\n",
    "        text_x1, text_y1, text_x2, text_y2 = draw.textbbox((x1,y1),bidi_ar_text, font=font)         #get the text size\n",
    "\n",
    "        height = (text_y2 - text_y1) + 5\n",
    "        width = (text_x2 - text_x1) + 5\n",
    "\n",
    "        draw.rectangle([(x1, y1-height), (x1+width, y1)], (0,255,0), width=2)\n",
    "        draw.text((x1+3 , y1-height), bidi_ar_text, font=font, fill=(0, 0, 0))\n",
    "\n",
    "        #english label\n",
    "        reshaped_en_label = label_to_en(label)\n",
    "        bidi_en_text = get_display(reshaped_en_label)\n",
    "        text1_x1, text1_y1, text1_x2, text1_y2 = draw.textbbox((x1,y1),bidi_en_text, font=font)\n",
    "\n",
    "        height1 = (text1_y2 - text1_y1) + 5\n",
    "        width1 = (text1_x2 - text1_x1) + 5\n",
    "\n",
    "        draw.rectangle([(x1, y2), (x1+width1, y2+height1)], (0,255,0), width=2)\n",
    "        draw.text((x1+3 , y2), bidi_en_text, font=font, fill=(0, 0, 0))\n",
    "\n",
    "    img = cv2.cvtColor(np.array(img_pil), cv2.COLOR_RGB2BGR)\n",
    "\n",
    "    #os.makedirs(save_dir, exist_ok=True)\n",
    "\n",
    "    # Save the image with bounding boxes and labels\n",
    "    save_path = os.path.join(save_dir, os.path.basename(filename))\n",
    "    cv2.imwrite(save_path, img)\n",
    "\n",
    "    return save_path, img\n",
    "\n",
    "\n",
    "def save_pred(image_name, bbox, labels, save_dir):\n",
    "\n",
    "    txt_folder = os.path.join(save_dir, \"labels\")\n",
    "    os.makedirs(txt_folder, exist_ok=True)\n",
    "    txt_file = os.path.join(txt_folder, f\"{image_name}.txt\")\n",
    "\n",
    "    with open (txt_file, 'w') as txt:\n",
    "        for box, label in zip(bbox, labels):\n",
    "            x1, y1, x2, y2 = box\n",
    "            en_label = label_to_en(label)\n",
    "            txt.write(f\"Label : {label}, Eng_label : {en_label}, Coordinates : {x1} {y1} {x2} {y2}\")\n",
    "    return txt_file\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "#multiple pred test\n",
    "\n",
    "\n",
    "def numberplate_recognition(source):\n",
    "\n",
    "    start = time.time()\n",
    "\n",
    "    model_path = \"models/yv8.pt\"\n",
    "    save_dir = \"runs\"\n",
    "    model = numplate_model\n",
    "\n",
    "    yolo_model = YOLO(model_path)\n",
    "    os.makedirs(save_dir, exist_ok=True)\n",
    "    \n",
    "    if os.path.isdir(source):\n",
    "        # Iterate over all files\n",
    "        for filename in os.listdir(source):\n",
    "            image_path = os.path.join(source, filename)\n",
    "            if os.path.isfile(image_path) and filename.lower().endswith(('.png', '.jpg', '.jpeg')):\n",
    "\n",
    "                bounding_boxes, labels, img = numberplate_detection(image_path,yolo_model,model)\n",
    "                fname = os.path.splitext(filename)[0]\n",
    "                savepath, img = bbox_and_labels(img,bounding_boxes,labels,save_dir,filename)\n",
    "\n",
    "                resized = imutils.resize(img, width=640)\n",
    "                cv2.imshow(\"output figure\",resized)\n",
    "                cv2.waitKey(0)\n",
    "                \n",
    "                textfile = save_pred(fname,bounding_boxes,labels,save_dir)\n",
    "                print(f\"predicted to {savepath} \\npredictions saved to {textfile}\")\n",
    "                \n",
    "    elif os.path.isfile(source):\n",
    "        bounding_boxes, labels, img = numberplate_detection(source,yolo_model,model)\n",
    "        fname = os.path.basename(source)\n",
    "        filename = os.path.splitext(fname)[0]\n",
    "        savepath, img = bbox_and_labels(img,bounding_boxes,labels,save_dir,fname)\n",
    "        resized = imutils.resize(img, width=640)\n",
    "        cv2.imshow(\"output figure\",resized)\n",
    "        cv2.waitKey(0)\n",
    "        textfile = save_pred(filename,bounding_boxes,labels,save_dir)\n",
    "        print(f\"predicted to {savepath} \\npredictions saved to {textfile}\")\n",
    "\n",
    "    else:\n",
    "        print(f\"Source path '{source}' is neither a valid directory nor a file.\")\n",
    "    cv2.destroyAllWindows()\n",
    "    end = time.time()\n",
    "    print(f\"Time taken: {end - start:.2f} seconds\")\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0034.jpg: 640x640 1 LicensePlate, 113.3ms\n",
      "Speed: 8.2ms preprocess, 113.3ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 73ms/step\n",
      "['د ط 1']\n",
      "predicted to runs\\0034.jpg \n",
      "predictions saved to runs\\labels\\0034.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0036.jpg: 640x480 1 LicensePlate, 87.3ms\n",
      "Speed: 0.0ms preprocess, 87.3ms inference, 0.5ms postprocess per image at shape (1, 3, 640, 480)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 61ms/step\n",
      "['د ف ق 5 3 4 9']\n",
      "predicted to runs\\0036.jpg \n",
      "predictions saved to runs\\labels\\0036.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0038.jpg: 640x640 1 LicensePlate, 100.2ms\n",
      "Speed: 4.8ms preprocess, 100.2ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 60ms/step\n",
      "['س 2']\n",
      "predicted to runs\\0038.jpg \n",
      "predictions saved to runs\\labels\\0038.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0074.jpg: 640x640 1 LicensePlate, 98.8ms\n",
      "Speed: 14.9ms preprocess, 98.8ms inference, 7.9ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 70ms/step\n",
      "['ج ن ى 7 7 7']\n",
      "predicted to runs\\0074.jpg \n",
      "predictions saved to runs\\labels\\0074.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0085.jpg: 640x512 1 LicensePlate, 87.6ms\n",
      "Speed: 10.0ms preprocess, 87.6ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 63ms/step\n",
      "['د ل ع 4 4 4 4']\n",
      "predicted to runs\\0085.jpg \n",
      "predictions saved to runs\\labels\\0085.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0134.jpg: 640x480 1 LicensePlate, 74.9ms\n",
      "Speed: 0.0ms preprocess, 74.9ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 480)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 60ms/step\n",
      "['ا ج ر 5 5 5']\n",
      "predicted to runs\\0134.jpg \n",
      "predictions saved to runs\\labels\\0134.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0137.jpg: 384x640 1 LicensePlate, 110.1ms\n",
      "Speed: 6.7ms preprocess, 110.1ms inference, 0.0ms postprocess per image at shape (1, 3, 384, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 68ms/step\n",
      "['ا ه م 4 4 4']\n",
      "predicted to runs\\0137.jpg \n",
      "predictions saved to runs\\labels\\0137.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0155.jpg: 384x640 1 LicensePlate, 81.1ms\n",
      "Speed: 2.1ms preprocess, 81.1ms inference, 0.8ms postprocess per image at shape (1, 3, 384, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 72ms/step\n",
      "['د د د 6 6 6 6']\n",
      "predicted to runs\\0155.jpg \n",
      "predictions saved to runs\\labels\\0155.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0201.jpg: 640x640 1 LicensePlate, 115.5ms\n",
      "Speed: 6.3ms preprocess, 115.5ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 59ms/step\n",
      "['م و ن 6 6 6 6']\n",
      "predicted to runs\\0201.jpg \n",
      "predictions saved to runs\\labels\\0201.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0216.jpg: 448x640 1 LicensePlate, 57.8ms\n",
      "Speed: 0.0ms preprocess, 57.8ms inference, 0.0ms postprocess per image at shape (1, 3, 448, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 59ms/step\n",
      "['ط ل ق 6 6 6']\n",
      "predicted to runs\\0216.jpg \n",
      "predictions saved to runs\\labels\\0216.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0240.jpg: 480x640 1 LicensePlate, 91.5ms\n",
      "Speed: 1.5ms preprocess, 91.5ms inference, 0.0ms postprocess per image at shape (1, 3, 480, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 74ms/step\n",
      "['م 1']\n",
      "predicted to runs\\0240.jpg \n",
      "predictions saved to runs\\labels\\0240.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0241.jpg: 640x640 1 LicensePlate, 80.7ms\n",
      "Speed: 4.0ms preprocess, 80.7ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 52ms/step\n",
      "['ى 1']\n",
      "predicted to runs\\0241.jpg \n",
      "predictions saved to runs\\labels\\0241.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0253.jpg: 640x640 1 LicensePlate, 99.2ms\n",
      "Speed: 0.0ms preprocess, 99.2ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 65ms/step\n",
      "['ص ص 2 2 2 2']\n",
      "predicted to runs\\0253.jpg \n",
      "predictions saved to runs\\labels\\0253.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0272.jpg: 512x640 1 LicensePlate, 112.7ms\n",
      "Speed: 0.0ms preprocess, 112.7ms inference, 0.0ms postprocess per image at shape (1, 3, 512, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 75ms/step\n",
      "['د ف ل 6 6 6']\n",
      "predicted to runs\\0272.jpg \n",
      "predictions saved to runs\\labels\\0272.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0292.jpg: 640x512 1 LicensePlate, 155.6ms\n",
      "Speed: 4.5ms preprocess, 155.6ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 81ms/step\n",
      "['ا ا 1 1 1']\n",
      "predicted to runs\\0292.jpg \n",
      "predictions saved to runs\\labels\\0292.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0293.jpg: 352x640 1 LicensePlate, 116.7ms\n",
      "Speed: 0.0ms preprocess, 116.7ms inference, 0.0ms postprocess per image at shape (1, 3, 352, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 91ms/step\n",
      "['ن و ر 1']\n",
      "predicted to runs\\0293.jpg \n",
      "predictions saved to runs\\labels\\0293.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0294.jpg: 352x640 1 LicensePlate, 103.6ms\n",
      "Speed: 13.0ms preprocess, 103.6ms inference, 0.0ms postprocess per image at shape (1, 3, 352, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 70ms/step\n",
      "['ف ر س 1 1 1']\n",
      "predicted to runs\\0294.jpg \n",
      "predictions saved to runs\\labels\\0294.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0314.jpg: 640x512 1 LicensePlate, 94.3ms\n",
      "Speed: 3.6ms preprocess, 94.3ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 69ms/step\n",
      "['س ع د 1 1 1 1']\n",
      "predicted to runs\\0314.jpg \n",
      "predictions saved to runs\\labels\\0314.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0329.jpg: 640x512 1 LicensePlate, 91.6ms\n",
      "Speed: 4.8ms preprocess, 91.6ms inference, 8.1ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step\n",
      "['ج د ع 1 1 1']\n",
      "predicted to runs\\0329.jpg \n",
      "predictions saved to runs\\labels\\0329.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0334.jpg: 384x640 1 LicensePlate, 77.2ms\n",
      "Speed: 0.0ms preprocess, 77.2ms inference, 0.0ms postprocess per image at shape (1, 3, 384, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 64ms/step\n",
      "['م ه م 2 2 2']\n",
      "predicted to runs\\0334.jpg \n",
      "predictions saved to runs\\labels\\0334.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0346.jpg: 608x640 1 LicensePlate, 173.7ms\n",
      "Speed: 0.5ms preprocess, 173.7ms inference, 0.0ms postprocess per image at shape (1, 3, 608, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step\n",
      "['د د 7 7 7 7']\n",
      "predicted to runs\\0346.jpg \n",
      "predictions saved to runs\\labels\\0346.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0368.jpg: 640x512 1 LicensePlate, 87.8ms\n",
      "Speed: 0.0ms preprocess, 87.8ms inference, 8.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step\n",
      "['و و و 6 6 6']\n",
      "predicted to runs\\0368.jpg \n",
      "predictions saved to runs\\labels\\0368.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0389.jpg: 640x640 1 LicensePlate, 134.9ms\n",
      "Speed: 0.0ms preprocess, 134.9ms inference, 8.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 68ms/step\n",
      "['ص ق ر 8 8 8 8']\n",
      "predicted to runs\\0389.jpg \n",
      "predictions saved to runs\\labels\\0389.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0403.jpg: 640x512 1 LicensePlate, 143.8ms\n",
      "Speed: 0.0ms preprocess, 143.8ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 79ms/step\n",
      "['ى ى 5 5 5 5']\n",
      "predicted to runs\\0403.jpg \n",
      "predictions saved to runs\\labels\\0403.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0429.jpg: 640x512 1 LicensePlate, 143.8ms\n",
      "Speed: 3.8ms preprocess, 143.8ms inference, 6.3ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 98ms/step\n",
      "['ا ا ا 4 4 4 4']\n",
      "predicted to runs\\0429.jpg \n",
      "predictions saved to runs\\labels\\0429.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0432.jpg: 640x512 1 LicensePlate, 102.1ms\n",
      "Speed: 0.0ms preprocess, 102.1ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step\n",
      "['ا ا 1']\n",
      "predicted to runs\\0432.jpg \n",
      "predictions saved to runs\\labels\\0432.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0507.jpg: 480x640 1 LicensePlate, 143.9ms\n",
      "Speed: 0.0ms preprocess, 143.9ms inference, 0.0ms postprocess per image at shape (1, 3, 480, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 79ms/step\n",
      "['ر 1']\n",
      "predicted to runs\\0507.jpg \n",
      "predictions saved to runs\\labels\\0507.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0515.jpg: 640x512 1 LicensePlate, 97.5ms\n",
      "Speed: 15.6ms preprocess, 97.5ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 66ms/step\n",
      "['م ط ر 1']\n",
      "predicted to runs\\0515.jpg \n",
      "predictions saved to runs\\labels\\0515.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0552.jpg: 640x512 1 LicensePlate, 96.0ms\n",
      "Speed: 2.3ms preprocess, 96.0ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 64ms/step\n",
      "['ن س ر 9 9 9 9']\n",
      "predicted to runs\\0552.jpg \n",
      "predictions saved to runs\\labels\\0552.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0554.jpg: 640x512 1 LicensePlate, 85.2ms\n",
      "Speed: 0.0ms preprocess, 85.2ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 55ms/step\n",
      "['ع ف 2']\n",
      "predicted to runs\\0554.jpg \n",
      "predictions saved to runs\\labels\\0554.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0565.jpg: 640x512 2 LicensePlates, 136.2ms\n",
      "Speed: 0.9ms preprocess, 136.2ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 83ms/step\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 72ms/step\n",
      "['ا ن س 3 8 4', 'ا ن س 2 2 2']\n",
      "predicted to runs\\0565.jpg \n",
      "predictions saved to runs\\labels\\0565.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0659.jpg: 640x512 1 LicensePlate, 85.8ms\n",
      "Speed: 0.0ms preprocess, 85.8ms inference, 15.7ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 65ms/step\n",
      "['ف 1 1']\n",
      "predicted to runs\\0659.jpg \n",
      "predictions saved to runs\\labels\\0659.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0668.jpg: 640x512 1 LicensePlate, 110.9ms\n",
      "Speed: 0.0ms preprocess, 110.9ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 65ms/step\n",
      "['ن ا ر 1 1 1 1']\n",
      "predicted to runs\\0668.jpg \n",
      "predictions saved to runs\\labels\\0668.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0681.jpg: 640x608 1 LicensePlate, 164.4ms\n",
      "Speed: 0.0ms preprocess, 164.4ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 608)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 70ms/step\n",
      "['م ى ر 1']\n",
      "predicted to runs\\0681.jpg \n",
      "predictions saved to runs\\labels\\0681.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0685.jpg: 640x640 1 LicensePlate, 100.7ms\n",
      "Speed: 5.4ms preprocess, 100.7ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 68ms/step\n",
      "['ج ب ل 7 7 7']\n",
      "predicted to runs\\0685.jpg \n",
      "predictions saved to runs\\labels\\0685.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0717.jpg: 640x512 1 LicensePlate, 129.5ms\n",
      "Speed: 0.0ms preprocess, 129.5ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 88ms/step\n",
      "['ا ط 6 6 6 6']\n",
      "predicted to runs\\0717.jpg \n",
      "predictions saved to runs\\labels\\0717.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0730.jpg: 640x512 1 LicensePlate, 97.2ms\n",
      "Speed: 0.0ms preprocess, 97.2ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 63ms/step\n",
      "['ا ر 4 4 4 4']\n",
      "predicted to runs\\0730.jpg \n",
      "predictions saved to runs\\labels\\0730.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0740.jpg: 640x640 1 LicensePlate, 169.4ms\n",
      "Speed: 2.1ms preprocess, 169.4ms inference, 2.6ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 73ms/step\n",
      "['ا ع 1']\n",
      "predicted to runs\\0740.jpg \n",
      "predictions saved to runs\\labels\\0740.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0780.jpg: 640x640 1 LicensePlate, 100.8ms\n",
      "Speed: 8.3ms preprocess, 100.8ms inference, 0.5ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step\n",
      "['م م م 4 4 4 4']\n",
      "predicted to runs\\0780.jpg \n",
      "predictions saved to runs\\labels\\0780.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0878.jpg: 640x640 1 LicensePlate, 148.0ms\n",
      "Speed: 0.0ms preprocess, 148.0ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 72ms/step\n",
      "['س س س 7']\n",
      "predicted to runs\\0878.jpg \n",
      "predictions saved to runs\\labels\\0878.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0893.jpg: 640x640 1 LicensePlate, 166.1ms\n",
      "Speed: 6.5ms preprocess, 166.1ms inference, 0.8ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 75ms/step\n",
      "['ا ل ج 7 7 7 7']\n",
      "predicted to runs\\0893.jpg \n",
      "predictions saved to runs\\labels\\0893.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0897.jpg: 640x608 1 LicensePlate, 164.3ms\n",
      "Speed: 6.7ms preprocess, 164.3ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 608)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 74ms/step\n",
      "['م م 2']\n",
      "predicted to runs\\0897.jpg \n",
      "predictions saved to runs\\labels\\0897.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0928.jpg: 640x640 1 LicensePlate, 172.2ms\n",
      "Speed: 4.0ms preprocess, 172.2ms inference, 5.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 73ms/step\n",
      "['ن ج م 6 6 6 6']\n",
      "predicted to runs\\0928.jpg \n",
      "predictions saved to runs\\labels\\0928.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0957.jpg: 608x640 1 LicensePlate, 113.9ms\n",
      "Speed: 0.0ms preprocess, 113.9ms inference, 1.9ms postprocess per image at shape (1, 3, 608, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 69ms/step\n",
      "['د د د 7 7 7']\n",
      "predicted to runs\\0957.jpg \n",
      "predictions saved to runs\\labels\\0957.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\0999.jpg: 640x640 1 LicensePlate, 178.9ms\n",
      "Speed: 9.2ms preprocess, 178.9ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 73ms/step\n",
      "['ز 1']\n",
      "predicted to runs\\0999.jpg \n",
      "predictions saved to runs\\labels\\0999.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1049.jpg: 640x640 2 LicensePlates, 137.7ms\n",
      "Speed: 11.1ms preprocess, 137.7ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 83ms/step\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 89ms/step\n",
      "['س س س 1 1 1', 'س ر ع 6 5 8 4']\n",
      "predicted to runs\\1049.jpg \n",
      "predictions saved to runs\\labels\\1049.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1059.jpg: 640x512 1 LicensePlate, 104.2ms\n",
      "Speed: 0.0ms preprocess, 104.2ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 72ms/step\n",
      "['ن س ر 6 6 6 6']\n",
      "predicted to runs\\1059.jpg \n",
      "predictions saved to runs\\labels\\1059.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1194.jpg: 480x640 1 LicensePlate, 140.4ms\n",
      "Speed: 8.4ms preprocess, 140.4ms inference, 1.1ms postprocess per image at shape (1, 3, 480, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 70ms/step\n",
      "['ر ر ر 7 7 7 7']\n",
      "predicted to runs\\1194.jpg \n",
      "predictions saved to runs\\labels\\1194.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1214.jpg: 544x640 1 LicensePlate, 100.7ms\n",
      "Speed: 1.6ms preprocess, 100.7ms inference, 0.0ms postprocess per image at shape (1, 3, 544, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 68ms/step\n",
      "['ا ا 4 4 4 4']\n",
      "predicted to runs\\1214.jpg \n",
      "predictions saved to runs\\labels\\1214.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1252.jpg: 640x512 1 LicensePlate, 112.9ms\n",
      "Speed: 2.2ms preprocess, 112.9ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 512)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 84ms/step\n",
      "['ا 8']\n",
      "predicted to runs\\1252.jpg \n",
      "predictions saved to runs\\labels\\1252.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1298.jpg: 480x640 2 LicensePlates, 129.4ms\n",
      "Speed: 0.0ms preprocess, 129.4ms inference, 0.0ms postprocess per image at shape (1, 3, 480, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 81ms/step\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 69ms/step\n",
      "['ج ب ل 1 1 1 1', 'ا س د 4 4 4 4']\n",
      "predicted to runs\\1298.jpg \n",
      "predictions saved to runs\\labels\\1298.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1345.jpg: 480x640 1 LicensePlate, 81.3ms\n",
      "Speed: 0.0ms preprocess, 81.3ms inference, 0.0ms postprocess per image at shape (1, 3, 480, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step\n",
      "['ى ى ى 8 8 8 8']\n",
      "predicted to runs\\1345.jpg \n",
      "predictions saved to runs\\labels\\1345.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1483.jpg: 640x640 1 LicensePlate, 145.7ms\n",
      "Speed: 8.2ms preprocess, 145.7ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 70ms/step\n",
      "['ن ا 1 1 1 1']\n",
      "predicted to runs\\1483.jpg \n",
      "predictions saved to runs\\labels\\1483.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1595.jpg: 384x640 2 LicensePlates, 116.0ms\n",
      "Speed: 0.0ms preprocess, 116.0ms inference, 0.0ms postprocess per image at shape (1, 3, 384, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 95ms/step\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 73ms/step\n",
      "['ن ج ا 6 6 6 6', 'ن ج ا 1 1 1']\n",
      "predicted to runs\\1595.jpg \n",
      "predictions saved to runs\\labels\\1595.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1666.jpg: 480x640 3 LicensePlates, 93.2ms\n",
      "Speed: 0.0ms preprocess, 93.2ms inference, 0.0ms postprocess per image at shape (1, 3, 480, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 66ms/step\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 53ms/step\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 64ms/step\n",
      "['ل ج م 5 5 5 5', 'ى ى ى 3 2 3 3', 'ج ن 1']\n",
      "predicted to runs\\1666.jpg \n",
      "predictions saved to runs\\labels\\1666.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1715.jpg: 480x640 2 LicensePlates, 112.5ms\n",
      "Speed: 2.7ms preprocess, 112.5ms inference, 4.5ms postprocess per image at shape (1, 3, 480, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 78ms/step\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 61ms/step\n",
      "['ب س م 1 1 1', 'ق ل ق 8 8 8 8']\n",
      "predicted to runs\\1715.jpg \n",
      "predictions saved to runs\\labels\\1715.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1849.jpg: 480x640 1 LicensePlate, 74.2ms\n",
      "Speed: 12.2ms preprocess, 74.2ms inference, 5.5ms postprocess per image at shape (1, 3, 480, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 69ms/step\n",
      "['و ق 3 9 3 2']\n",
      "predicted to runs\\1849.jpg \n",
      "predictions saved to runs\\labels\\1849.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1852.jpg: 640x480 1 LicensePlate, 83.9ms\n",
      "Speed: 14.3ms preprocess, 83.9ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 480)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step\n",
      "['د ق 5 4 7 3']\n",
      "predicted to runs\\1852.jpg \n",
      "predictions saved to runs\\labels\\1852.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\1861.jpg: 640x480 1 LicensePlate, 143.0ms\n",
      "Speed: 0.0ms preprocess, 143.0ms inference, 0.0ms postprocess per image at shape (1, 3, 640, 480)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 91ms/step\n",
      "['ر ب ب 5 6 2']\n",
      "predicted to runs\\1861.jpg \n",
      "predictions saved to runs\\labels\\1861.txt\n",
      "\n",
      "image 1/1 x:\\DL\\Numberplate\\data\\tset\\2015.jpg: 480x640 2 LicensePlates, 96.8ms\n",
      "Speed: 3.6ms preprocess, 96.8ms inference, 0.0ms postprocess per image at shape (1, 3, 480, 640)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 72ms/step\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 63ms/step\n",
      "['ب م 1 1 1 1', 'ب م 6 6 6 6']\n",
      "predicted to runs\\2015.jpg \n",
      "predictions saved to runs\\labels\\2015.txt\n",
      "Time taken: 36.99 seconds\n"
     ]
    }
   ],
   "source": [
    "test_set = \"data/tset\"\n",
    "\n",
    "numberplate_recognition(test_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "v8OPS",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}