Skip to content

Commit

Permalink
first commit
Browse files Browse the repository at this point in the history
  • Loading branch information
@zishansami102
zishansami102 committed Feb 8, 2018
1 parent f4900f0 commit fc11b8d
Show file tree
Hide file tree
Showing 11 changed files with 1,062 additions and 1 deletion.
Binary file added 1.pdf
View file
Binary file not shown.
66 changes: 65 additions & 1 deletion README.md
View file
Original file line number Diff line number Diff line change
@@ -1 +1,65 @@
# First-Impression
# First-Impression

This is the solution to the poblem "First Impressions V2" given in CVPR'17, ECCV '16 & ICPR '16 and this piece of code is the implementation of the [paper](https://cs.nju.edu.cn/wujx/paper/eccvw16_APA.pdf) which is the winner of ECCV 2016

This problem is a challenge on “first impressions”, in which participants will develop solutions for recognizing personality traits of users in short video sequences. They have made available a large newly collected data set sponsored by Microsoft of at least 10,000 15-second videos collected from YouTube, annotated with personality traits by AMT workers.

The traits to be recognized will correspond to the “big five” personality traits used in psychology and well known of hiring managers using standardized personality profiling:
* Extroversion
* Agreeableness
* Conscientiousness
* Neuroticism
* Openness to experience.

As is known, the first impression made is highly important in many contexts, such as human resourcing or job interviews. This work could become very relevant to training young people to present themselves better by changing their behavior in simple ways. The participants who obtain the best results in the challenge will be invited to submit a paper to the workshop.


## Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

### Prerequisites

* [Python2](https://www.python.org/download/releases/2.7.2/) - Python version 2.7.2
* [Numpy](http://www.numpy.org/) - Multidimensioanl Mathematical Computing
* [Tensorflow](https://www.tensorflow.org/) - Deep Learning python module
* [Pandas](https://pandas.pydata.org/) -
* [Cha-Learn Dataset](http://chalearnlap.cvc.uab.es/dataset/24/description/) - Dataset for this problem
* [Pretrained VGG16 Model](https://www.cs.toronto.edu/~frossard/vgg16/vgg16_weights.npz) Pretrained Vgg16 mode

### Installing

* Clone the repository

```
git clone https://github.com/zishansami102/CNN-from-Scratch
```

* Downlad the training dataset and copy that to this folder and extract it into a new /data directory with all 75 zip files as it is

* [Download](https://www.cs.toronto.edu/~frossard/vgg16/vgg16_weights.npz) Pretrained Vgg16 model and move it to the root directory


* Run the vidToimg.py file to scrape the images from the videos and save it to a new ImageData directory

```
python VidToimg.py
```


* If succesfully completed then run the savetfreco.py file to form a data pipeline by saving the all the images into train500.tfrecords file to load it later during training

```
python savetfreco.py
```

* Start the training by running the following command

```
python train.py
## Acknowledgments
* [paper](https://cs.nju.edu.cn/wujx/paper/eccvw16_APA.pdf) - Implemented paper
* [TfRecord Data Pipeline](http://machinelearninguru.com/deep_learning/data_preparation/tfrecord/tfrecord.html#read) - Used to make data pipeline
251 changes: 251 additions & 0 deletions dan.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,251 @@
########################################################################################
# Davi Frossard, 2016 #
# VGG16 implementation in TensorFlow #
# Details: #
# http://www.cs.toronto.edu/~frossard/post/vgg16/ #
# #
# Model from https://gist.github.com/ksimonyan/211839e770f7b538e2d8#file-readme-md #
# Weights from Caffe converted using https://github.com/ethereon/caffe-tensorflow #
########################################################################################

import tensorflow as tf
import numpy as np
import warnings


warnings.filterwarnings("ignore")




class DAN:
def __init__(self, imgs, weights=None, sess=None):
self.imgs = imgs
self.convlayers()
self.dan_part()
if weights is not None and sess is not None:
self.load_weights(weights, sess)
self.output = tf.nn.sigmoid(self.reg_head)



def convlayers(self):
self.parameters = []

# zero-mean input
with tf.name_scope('preprocess') as scope:
mean = tf.constant([123.68, 116.779, 103.939], dtype=tf.float32, shape=[1, 1, 1, 3], name='img_mean')
images = self.imgs-mean

# conv1_1
with tf.name_scope('conv1_1') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 3, 64], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(images, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[64], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv1_1 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# conv1_2
with tf.name_scope('conv1_2') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 64, 64], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.conv1_1, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[64], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv1_2 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# pool1
self.pool1 = tf.nn.max_pool(self.conv1_2,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME',
name='pool1')

# conv2_1
with tf.name_scope('conv2_1') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 64, 128], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.pool1, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[128], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv2_1 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# conv2_2
with tf.name_scope('conv2_2') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 128, 128], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.conv2_1, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[128], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv2_2 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# pool2
self.pool2 = tf.nn.max_pool(self.conv2_2,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME',
name='pool2')

# conv3_1
with tf.name_scope('conv3_1') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 128, 256], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.pool2, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[256], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv3_1 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# conv3_2
with tf.name_scope('conv3_2') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 256, 256], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.conv3_1, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[256], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv3_2 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# conv3_3
with tf.name_scope('conv3_3') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 256, 256], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.conv3_2, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[256], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv3_3 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# pool3
self.pool3 = tf.nn.max_pool(self.conv3_3,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME',
name='pool3')

# conv4_1
with tf.name_scope('conv4_1') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 256, 512], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.pool3, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv4_1 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# conv4_2
with tf.name_scope('conv4_2') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 512, 512], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.conv4_1, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv4_2 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# conv4_3
with tf.name_scope('conv4_3') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 512, 512], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.conv4_2, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv4_3 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# pool4
self.pool4 = tf.nn.max_pool(self.conv4_3,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME',
name='pool4')

# conv5_1
with tf.name_scope('conv5_1') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 512, 512], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.pool4, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv5_1 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# conv5_2
with tf.name_scope('conv5_2') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 512, 512], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.conv5_1, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv5_2 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# conv5_3
with tf.name_scope('conv5_3') as scope:
kernel = tf.Variable(tf.truncated_normal([3, 3, 512, 512], dtype=tf.float32,
stddev=1e-1), name='weights')
conv = tf.nn.conv2d(self.conv5_2, kernel, [1, 1, 1, 1], padding='SAME')
biases = tf.Variable(tf.constant(0.0, shape=[512], dtype=tf.float32),
trainable=True, name='biases')
out = tf.nn.bias_add(conv, biases)
self.conv5_3 = tf.nn.relu(out, name=scope)
self.parameters += [kernel, biases]

# MaxPool5
self.maxpool5 = tf.nn.max_pool(self.conv5_3,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME',
name='maxpool5')

# AvgPool5
self.avgpool5 = tf.nn.avg_pool(self.conv5_3,
ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1],
padding='SAME',
name='avgpool5')

def dan_part(self):

# fc1
with tf.name_scope('reg_head') as scope:
shape = 2*int(np.prod(self.maxpool5.get_shape()[1:]))
# shape = int(np.prod(self.pool5.get_shape()[1:]))
fc1w = tf.Variable(tf.truncated_normal([shape, 5],
dtype=tf.float32,
stddev=1e-1), name='weights')
fc1b = tf.Variable(tf.constant(1.0, shape=[5], dtype=tf.float32),
trainable=True, name='biases')

maxpool5_flat = tf.nn.l2_normalize(tf.reshape(self.maxpool5, [-1, shape/2]), 1)
avgpool5_flat = tf.nn.l2_normalize(tf.reshape(self.avgpool5, [-1, shape/2]), 1)

self.concat = tf.concat([maxpool5_flat, avgpool5_flat], 1)
self.reg_head = tf.nn.bias_add(tf.matmul(self.concat, fc1w), fc1b)
self.parameters += [fc1w, fc1b]


def load_weights(self, weight_file, sess):
weights = np.load(weight_file)
keys = sorted(weights.keys())
for i, k in enumerate(keys):
if i==len(self.parameters)-2:
break
sess.run(self.parameters[i].assign(weights[k]))
Binary file added dan.pyc
View file
Binary file not shown.
46 changes: 46 additions & 0 deletions remtime.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,46 @@
import tensorflow as tf

def printTime(remtime):
hrs = int(remtime)/3600
mins = int((remtime/60-hrs*60))
secs = int(remtime-mins*60-hrs*3600)
print(str(hrs)+"Hrs "+str(mins)+"Mins "+str(secs)+"Secs remaining..")

def dataBatch(data_path, BATCH_SIZE, N_EPOCHS=1):
reader = tf.TFRecordReader()
filename_queue = tf.train.string_input_producer([data_path], num_epochs=N_EPOCHS)
_, serialized_example = reader.read(filename_queue)
# Decode the record read by the reader
feature = {'train/image': tf.FixedLenFeature([], tf.string), 'train/label': tf.FixedLenFeature([], tf.string)}
features = tf.parse_single_example(serialized_example, features=feature)
# Convert the image data from string back to the numbers
image = tf.decode_raw(features['train/image'], tf.float32)
label = tf.decode_raw(features['train/label'], tf.float32)
# Reshape image data into the original shape
image = tf.reshape(image, [224, 224, 3])
label = tf.reshape(label, [5])

images, labels = tf.train.shuffle_batch([image, label], batch_size=BATCH_SIZE, capacity=100, min_after_dequeue=BATCH_SIZE, allow_smaller_final_batch=True)
return images, labels



# VID_BATCH = 100


# df = pd.read_csv('small_train_sample.csv')
# NUM_VID = len(df)
# images=np.empty((NUM_VID*100, 224, 224, 3))
# traits=np.zeros((NUM_VID*100, 5))

# for i in range(NUM_VID):
# filelist=glob.glob('ImageData/trainingData/'+(df['VideoName'].iloc[i]).split('.mp4')[0]+'/*.jpg')
# print(str(i)+" "+ str(len(filelist))+" "+df['VideoName'].iloc[i])
# images[i*100:(i+1)*100] = np.array([np.array(Image.open(fname).resize((224,224), Image.ANTIALIAS)) for fname in filelist])
# traits[i*100:(i+1)*100]=np.array(df.drop(['VideoName'], 1, inplace=False).iloc[i])


# ind = [i for i in range(NUM_IMAGES)]

# print(images.shape)
# print(traits.shape)
Binary file added remtime.pyc
View file
Binary file not shown.
Loading

0 comments on commit fc11b8d

Please sign in to comment.