使用 cv2 收集图像数据时出现问题

Question

I'm trying to get the image data from some cat and dog images with cv2 for a machine learning project in python and append them all to a training_data list. But it is just printing None when printing the list at the end.

At first I thought that maybe I needed to convert the data to a numpy array but nothing seems to work and I'm really not understanding what's wrong.

CATEGORIES = ["Dog", "Cat"]     # 0=dog, 1=cat
IMG_SIZE = 50

training_data = []

def create_training_data():
    i = 0
    for category in CATEGORIES:
        path = os.path.join(DATADIR, category)  # path to cats or dogs dir
        class_num = CATEGORIES.index(category)
        for img in os.listdir(path):
            i += 1
            try:
                img_array = cv2.imread(os.path.join(path, img), cv2.IMREAD_GRAYSCALE)
                new_array = cv2.resize(img_array, (IMG_SIZE, IMG_SIZE))
                training_data.append([new_array, class_num])
            except Exception as e:
                print("general exception", e, os.path.join(path, img))
            if i % 1000 == 0:
                print(i, '...')
    print('\n\nfinished...\n\n')

create_training_data()

training_data = random.shuffle(training_data)

print(training_data)

X = []
y = []

for features, label in training_data:
    X.append(features)
    y.append(label)

X = np.array(X).reshape(-1, IMG_SIZE, IMG_SIZE, 1)

I think it has to do with the resizing because it is throwing a lot of these exceptions:

general exception OpenCV(4.1.1) C:...\opencv\modules\imgproc\src\resize.cpp:3720: error: (-215:Assertion failed).ssize:empty() in function 'cv::resize' D..../cats_and_dogs/PetImages\Cat\9565.jpg

It also prints this error:

Traceback (most recent call last):

File "D:/Python/tensorflow/tutorial/cats_and_dogs.py", line 44, in for features, label in training_data: TypeError: 'NoneType' object is not iterable

Corrupt JPEG data: 399 extraneous bytes before marker 0xd9

Corrupt JPEG data: 226 extraneous bytes before marker 0xd9

Corrupt JPEG data: 162 extraneous bytes before marker 0xd9

Warning: unknown JFIF revision number 0.00

...

Answer 1

After researching your errors I found this thread that may be relevant to the problem you are having.

In short, the thread notes a problem with the Oxford-IIIT Pet Dataset. In that dataset some of the images end in a '.jpg' extension but are actually '.png's. Verify where you got your images from.

It appears you are having a similar/the same problem. The mismatch between the file extension and the actual image data may lead to openCV not being able to read in the data correctly.

Edit: 2019 Oct. 14

Additionally, random.shuffle() operates in place and therefore returns None . Change training_data = random.shuffle(training_data) to random.shuffle(training_data)