# put image in 4D tensor of shape (1, 3, height, width) img_ = img.swapaxes(0, 2).swapaxes(1, 2).reshape(1, 3, img_h, img_w) filtered_img = f(img_) # plot original image and first and second components of output pylab.subplot(1, 3, 1); pylab.axis('off'); pylab.imshow(img) pylab.gray(); # recall that the convOp output (filtered image) is actually a "minibatch", # of size 1 here, so we take index 0 in the first dimension: pylab.subplot(1, 3, 2); pylab.axis('off'); pylab.imshow(filtered_img[0, 0, :, :]) pylab.subplot(1, 3, 3); pylab.axis('off'); pylab.imshow(filtered_img[0, 1, :, :]) # filter_img.shape = (1,2,img_h,img_w) #print type(filtered_img),filtered_img.shape #print type(filtered_img[0,0,:,:]),filtered_img[0,0,:,:].shape #print type(img),img.shape #pylab.show() ff = filtered_img[0,1,:,:] tmp_img = ff.reshape(ff.shape[0]*ff.shape[1]) #print tmp_img.shape tu = [] tu.append(sign) for i in tmp_img: tu.append(i) #print type(tu),len(tu) train_list.append(tu) print start,end tdtf.write_content_to_csv(train_list,DataHome + "test1.csv") print start,end
if end >= len(piclist): end = len(piclist) for i in range(start,end): #len(piclist)): img_route = piclist[i] img_route_list = img_route.split(".") breed = 1 if img_route_list[0] == "cat": breed = 0 else: breed = 1 img = Image.open(open(DataHome + img_data_dir + img_route)) img_w , img_h = img.size #if (img_w < 250) | (img_h < 250) : # continue #print "resizing" img = img.resize((250,250),Image.ANTIALIAS) features = get_feature(img) breed_list = [breed] img_info = breed_list + features #print type(img_info),len(img_info) data_set.append(img_info) print "writing from %d to %d " % (start,end) #print data_set tdtf.write_content_to_csv(data_set,DataHome + data_csv_file) #print len(features) #print features #print features == features #feature_file = open('color_feature.cPickle','w') #cPickle.dump(features,feature_file)
pylab.subplot(1, 3, 1) pylab.axis('off') pylab.imshow(img) pylab.gray() # recall that the convOp output (filtered image) is actually a "minibatch", # of size 1 here, so we take index 0 in the first dimension: pylab.subplot(1, 3, 2) pylab.axis('off') pylab.imshow(filtered_img[0, 0, :, :]) pylab.subplot(1, 3, 3) pylab.axis('off') pylab.imshow(filtered_img[0, 1, :, :]) # filter_img.shape = (1,2,img_h,img_w) #print type(filtered_img),filtered_img.shape #print type(filtered_img[0,0,:,:]),filtered_img[0,0,:,:].shape #print type(img),img.shape #pylab.show() ff = filtered_img[0, 1, :, :] tmp_img = ff.reshape(ff.shape[0] * ff.shape[1]) #print tmp_img.shape tu = [] tu.append(sign) for i in tmp_img: tu.append(i) #print type(tu),len(tu) train_list.append(tu) print start, end tdtf.write_content_to_csv(train_list, DataHome + "test1.csv") print start, end
#!/usr/bin/python import transform_data_to_format as tdtf test_l1 = [1,2,3,4,5,6] test_l2 = [1,2,3,4,5,6] test_l = (test_l1,test_l2) tdtf.write_content_to_csv(test_l,"../data/test.csv")
#!/usr/bin/python import transform_data_to_format as tdtf test_l1 = [1, 2, 3, 4, 5, 6] test_l2 = [1, 2, 3, 4, 5, 6] test_l = (test_l1, test_l2) tdtf.write_content_to_csv(test_l, "../data/test.csv")
bg = cv2.imread(DataHome + "bg/" + bg_name) bg = cv2.cvtColor(bg, cv2.COLOR_BGR2GRAY) h, w = bg.shape #show the gray img #cv2.imshow("w2",img) #cv2.waitKey(0) #reshape (h,w) to (h*w,) bg = bg.reshape(w * h) feature = [2] for f_v in bg: feature.append(f_v) features_list.append(feature) print len(features_list), len(features_list[0]), len(features_list[-1]) tdtf.write_content_to_csv(features_list, feature_filename) ''' train_index_list = random.sample(range(len(features_list)), len(features_list)/2 ) train_features_list = [] for i in train_index_list: train_features_list.append(features_list[i]) valid_features_list = [] for i in range(len(features_list)): if i in train_index_list: continue valid_features_list.append(features_list[i]) print len(train_features_list) print len(valid_features_list) tdtf.write_content_to_csv(train_features_list,train_feature_filename) '''
bg = cv2.imread(DataHome + "bg/" + bg_name) bg = cv2.cvtColor(bg, cv2.COLOR_BGR2GRAY) h, w = bg.shape #show the gray img #cv2.imshow("w2",img) #cv2.waitKey(0) #reshape (h,w) to (h*w,) bg=bg.reshape(w*h) feature= [2] for f_v in bg: feature.append(f_v) features_list.append(feature) print len(features_list),len(features_list[0]),len(features_list[-1]) tdtf.write_content_to_csv(features_list,feature_filename) ''' train_index_list = random.sample(range(len(features_list)), len(features_list)/2 ) train_features_list = [] for i in train_index_list: train_features_list.append(features_list[i]) valid_features_list = [] for i in range(len(features_list)): if i in train_index_list: continue valid_features_list.append(features_list[i]) print len(train_features_list) print len(valid_features_list) tdtf.write_content_to_csv(train_features_list,train_feature_filename) '''