def get_output_layer(model, layer_name, n): # get the symbolic outputs of each "key" layer (we gave them unique names). layer_dict = dict([(layer.name, layer) for layer in model.layers]) layer = layer_dict[layer_name].get_output_at(n) return layer #%% x_train = [] in_E = np.zeros((1, 6, im_size[0], im_size[1])) labels = np.zeros((1, 6, im_size[0], im_size[1])) print '\n---lead dataset Image \n' data_dir = './image/move_circle/' im_dir, dir_num = myl.ListDir(data_dir) print dir_num, im_dir im_sum = 0 for name in im_dir: print data_dir + name im = Image.open(data_dir + name) im = im.resize((im_size[1], im_size[0])) im = np.asarray(im.convert('RGB')) #im = ImageOps.grayscale(im) im = np.asarray(im) x_train.append(im) if len(x_train) > use_im_num: break print '\n---Image Nomarization & Reduction'