removed scipy and replaced by numpy

ai.py

@@ -1,9 +1,9 @@
 # Arda Mavi
 import os
 import platform
-import numpy as np
+import numpy as np, skimage, imageio, matplotlib
 from time import sleep
-from PIL import ImageGrab
+from PIL import ImageGrab, Image
 from game_control import *
 from predict import predict
 from game_control import *

create_dataset.py

@@ -1,13 +1,11 @@
-# Arda Mavi
 import os
 import sys
 import platform
-import numpy as np
+import numpy as np, skimage, imageio, matplotlib
 from time import sleep
-from PIL import ImageGrab
+from PIL import ImageGrab, Image
 from game_control import *
 from predict import predict
-from scipy.misc import imresize
 from game_control import get_id
 from get_dataset import save_img
 from multiprocessing import Process
@@ -18,7 +16,7 @@
 def get_screenshot():
     img = ImageGrab.grab()
     img = np.array(img)[:,:,:3] # Get first 3 channel from image as numpy array.
-    img = imresize(img, (150, 150, 3)).astype('float32')/255.
+    img = np.array(img, (150, 150, 3)).astype('float32')/255.
     return img
 
 def save_event_keyboard(data_path, event, key):

get_dataset.py

@@ -1,18 +1,17 @@
-# Arda Mavi
 import os
-import numpy as np
+import numpy as np, skimage, imageio, matplotlib
+from PIL import ImageGrab, Image
 from keras.utils import to_categorical
-from scipy.misc import imread, imresize, imsave
 from sklearn.model_selection import train_test_split
 
 def get_img(data_path):
     # Getting image array from path:
-    img = imread(data_path)
-    img = imresize(img, (150, 150, 3))
+    img = imageio.imread(data_path)
+    img = np.array(img, (150, 150, 3))
     return img
 
 def save_img(img, path):
-    imsave(path, img)
+    imageio.imwrite(path, img)
     return
 
 def get_dataset(dataset_path='Data/Train_Data'):

predict.py

@@ -1,8 +1,8 @@
 # Arda Mavi
-import numpy as np
-from scipy.misc import imresize
+import numpy as np, skimage, imageio, matplotlib
+from PIL import ImageGrab, Image
 
 def predict(model, X):
-    X = imresize(X, (150, 150, 3)).astype('float32')/255.
+    X = np.array(X, (150, 150, 3)).astype('float32')/255.
     Y = model.predict(X.reshape(1,150,150,3))
     return Y