#96_missing_fg-script and docs for fg-script

data_loader/fg-script.py

@@ -0,0 +1,94 @@
+import urllib.request
+import os
+import csv
+import cv2 as cv
+import random
+
+face_cascade = cv.CascadeClassifier('haarcascade_frontalface_default.xml')
+
+
+def createFolder(directory):
+    try:
+        if not os.path.exists(directory):
+            os.makedirs(directory)
+    except OSError:
+        print("Error: Creating directory.")
+
+def createClassificationFolders(directory):
+    createFolder(directory)
+    for x in range(101):
+        createFolder(directory + "/" + str(x))
+
+def load_images_from_folder(folder):
+    images = []
+    for subfolder in os.listdir(folder):
+        images.append(os.path.join(folder, subfolder))
+    return images
+
+if not os.path.exists("../data/FGNET.zip"):
+    print("Start downloading Zipfile")
+    urllib.request.urlretrieve("http://yanweifu.github.io/FG_NET_data/FGNET.zip", "../data/FGNET.zip")
+    print("Finished downloading Zipfile")
+else:
+    print("Zip already downloaded.")
+
+if not os.path.exists("../data/FGNET"):
+    print("Start unzipping")
+    import zipfile
+    with zipfile.ZipFile("../data/FGNET.zip", 'r') as zip_ref:
+        zip_ref.extractall("../data/FGNET")
+    print("Finished unzipping")
+else:
+    print("Already unzipped")
+
+
+age = []
+full_path = []
+full_path2 = load_images_from_folder("../data/FGNET/FGNET/images")
+delta = 5
+
+for a in full_path2:
+    b = a.split('\\')
+    full_path.append(b[1])
+    b = b[1].split('.')
+
+    c = b[0].split('A')
+    d = c[1].strip('a')
+    d = d.strip('b')
+    age.append(d)
+
+counter_training = 0
+counter_valid = 0
+counter_test = 0
+print("Printing Images to Classification Folders")
+
+for idx, val in enumerate(full_path):
+    img = cv.imread("../data/FGNET/FGNET/images/" + val)
+    faces = face_cascade.detectMultiScale(img, 1.8, 5)
+    for (x,y,w,h) in faces:
+        cip = img[y-delta:y+h+delta, x-delta:x+w+delta].copy()
+        try:
+            cip = cv.resize(cip, (224, 224))
+            i = random.random()
+            if i > 0.5:
+                counter_training = counter_training + 1
+                cv.imwrite("../data/LAP/Train/" + str(int(age[idx])) + "/" + val, cip)
+                print("Written: " + "../data/LAP/Train/" + str(int(age[idx])) + "/" + val)
+            elif i < 0.25:
+                counter_valid = counter_valid + 1
+                cv.imwrite("../data/LAP/Validation/" + str(int(age[idx])) + "/" + val, cip)
+                print("Written: " + "../data/LAP/Validation/" + str(int(age[idx])) + "/" + val)
+            else:
+                counter_test = counter_test + 1
+                cv.imwrite("../data/LAP/Test/" + str(int(age[idx])) + "/" + val, cip)
+                print("Written: " + "../data/LAP/Test/" + str(int(age[idx])) + "/" + val)
+        except:
+            print("Resize Fail")
+
+f = open("../data/img_counts_fg.txt", "w+")
+f.write("Train_img_count = " + str(counter_training) + "\n")
+f.write("Validation_img_count = " + str(counter_valid) + "\n")
+f.write("Test_img_count = " + str(counter_test) + "\n")
+f.close()
+
+print("FINISHED")

docs/fg-net.md

@@ -0,0 +1,17 @@
+#The "*fg-script.py*" file will -
+
+<pre><code>
+1. look if a file exists at *"../data/FGNET.zip"* exists - do nothing; else - download dataset from
+"*http://yanweifu.github.io/FG_NET_data/FGNET.zip"*
+
+2. look if an extracted file at *"../data/data/FGNET"* exists - do nothing; else - extract downloaded dataset
+
+3. have a function to create Classification folders in *"../data/LAP"* ranging from 0 to 100 (not used)
+
+4. open images and creates an array of tupels with the corresponding age value which is extracted from
+the image name
+
+5. write the images into the fitting Classification folder
+
+6. create a *"img_counts_fg.txt"* file with the amount of Training, Validiation and Test images
+</code></pre>

docs/lap_script.md

@@ -1,13 +1,16 @@
-The lap-script.py file will -
+#The "*lap-script.py*" file will -
+
 <pre><code>
-1. look if a file exists at *"../data/appa-real-release.zip"* exists - do nothing; else - download dataset 
+1. look if a file exists at *"../data/appa-real-release.zip"* exists - do nothing; else - download dataset from
+*"http://158.109.8.102/AppaRealAge/appa-real-release.zip*"
 
 2. look if an extracted file at *"../data/appa-real-release"* exists - do nothing; else - extract downloaded dataset
 
-3. create Classification folders in *"../data/LAP"* ranging
-   from 0 to 100
+3. create Classification folders in *"../data/LAP"* ranging from 0 to 100
 
 4. open images and creates an array of tupels with the corresponding age value from a csv file
 
-5. write the images into the fitting Classification folder 
-</code></pre>
+5. write the images into the fitting Classification folder
+
+6. create a *"img_counts.txt"* file with the amount of Training, Validiation and Test images
+</code></pre>