How to use a loop to save images in different folders in python-opencv

Question

I have more than 20 images that I want to classify based on pixel correlation. I am able to perform all the procedures but the problem is just saving the image into the corresponding class. Suppose that the pixel correlation values are[0.48, 0.20, 0.57, 0.53, 0.06, 0.52, 0.55, 0.57, 0.51, 0.49, ..., 0.25]

And I want to use each index of the above values based on some thresholds to classify the images into Normal, Abnormal and Ambiguous classes. With this snippet below, I achieved the following results however, The saved images are in an unsupported format

for filename in os.listdir(folder):
if filename.endswith('.jpg'):
    img = cv2.imread(os.path.join(folder, filename))
    gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
r = []
for correlation in range(len(correlation_matrices)):
    symmetry = correlation_matrices[correlation][0,1]
    r.append(symmetry)
#print(r)

threshold1 = 0.45
threshold2 = 0.35

for index in range(len(r)):

    
    if (rounded_value[index] >= threshold1):
        print('Normal Image')
        cv2.imwrite('./uniform_luminance/Normal' + str(index) + '.jpg', img[index])
    
    elif ((rounded_value[index] < threshold1) and rounded_value[index] >= threshold2):
        print('Ambiguous Image')
        cv2.imwrite('./ambiguous_luminance/Ambiguous' + str(index) + '.jpg', img[index])
    
    elif (rounded_value[index] < threshold2):
        print('Abnormal Image')
        cv2.imwrite('./non_uniform_luminance/Abnormal' + str(index) + '.jpg', img[index])

I achieved the following

The saved images are not similar to the input images however classified based on the thresholds

While the input images as follows

Iraklis Skepasianos · Accepted Answer · 2021-05-24 08:22:42Z

1

Are you asking how to correct the format of the saved images? Or why the saved images are different than your input ones?

In case you are asking the later:

Assuming your snippet is inside a loop that reads each image separately:

I believe your error is on img[index]. You are saving the image as a column, including all of the rows thus creating an image as shown. Keep in mind that an image is a 2d array and with [] you can specify which values you want to manipulate.

Image[columns, rows]

Since openvcv documentation says that imwrite takes parameters first the filename and second the image, change your code to:

cv2.imwrite('./uniform_luminance/Normal' + str(index) + '.jpg', img)

or

cv2.imwrite('./uniform_luminance/Normal' + str(index) + '.jpg', img[:,:])

img[:, :] the : is used to include all values.

answered May 24, 2021 at 8:22

Iraklis Skepasianos

688 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Kimwaga Makono Over a year ago

Thank you so much for your comment, I used index because when I printed the 3rd line in my code as print(img[0]) it returns a NumPy array

Iraklis Skepasianos Over a year ago

Okay. However, specifying which column to show ( img[0] ) will return all rows of that column. Is this what you want to achieve? I suggest trying: cv2.imshow('2D-Image', img) or cv2.imshow('2D-Image', img[:][:]) and compare it to the output of cv2.imshow('1D-Image', img[0]) so you can visually understand the difference. Please refer to: pythonexamples.org/python-opencv-imshow

Kimwaga Makono Over a year ago

Thank you so much for a clear clarification

Collectives™ on Stack Overflow

How to use a loop to save images in different folders in python-opencv

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related