TypeError：切片索引必须是整数或无或具有 __index__ 方法（​​Google Colab）

Question

I was trying to run the Step1_Data_RemoveNoise.ipynb notebook from microsoft/nestle-acne-assessment on Google Colab. When the image processing processed 100 images and above in the dataset, it started to throw an error TypeError: slice indices must be integers or None or have an __index__ method .

The error output:

No cheeks or forehead detected, output the original file levle1_128.jpg
Face not detected by landmarks model...
Left eye detected
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-7-93f89d71afa0> in <module>()
    165             if forehead_max_y - forehead_min_y >= 0.7 * eyes1[0, 3]:
    166                 forehead_ok = True
--> 167                 forehead_region = img[forehead_min_y:forehead_max_y, forehead_min_x: forehead_max_x, :]
    168                 forehead_region = cv2.cvtColor(forehead_region, cv2.COLOR_BGR2RGB)
    169                 forehead_file_name = join(croppedFaces_Dir, imageName+"_fh.jpg")

TypeError: slice indices must be integers or None or have an __index__ method

This is the levle1_128.jpg image. The error appeared here.

Answer 1

this occurs when we use a non-integer value for slicing.In this phase of coding

        ########################
        # Get the forehead patch
        ########################
        [right_brow_min_x, left_brow_max_x] = \
            [max(0, np.min(np.array(right_brow_landmarks[:,0]))), min(img_width, np.max(np.array(left_brow_landmarks[:,0])))]
        brow_min_y = min(np.min(np.array(right_brow_landmarks[:,1])),np.min(np.array(left_brow_landmarks[:,1])))
        forehead_x_min = right_brow_min_x # forehead starts at the left landmark of the right eye brow
        forehead_x_max = left_brow_max_x
        forehead_y_min = max(0, brow_min_y - forehead_height)
        forehead_y_max = min(brow_min_y, forehead_y_min + forehead_height)
        forehead_region = img[forehead_y_min:forehead_y_max, forehead_x_min:forehead_x_max, :]
        forehead_file_name = join(croppedFaces_Dir, imageName+"_fh.jpg")
        # BGR image needs to be converted to RGB before saving as image file
        forehead_region = cv2.cvtColor(forehead_region, cv2.COLOR_BGR2RGB)
        misc.imsave(forehead_file_name, forehead_region)

the line

forehead_y_min = max(0, brow_min_y - forehead_height)
            forehead_y_max = min(brow_min_y, forehead_y_min + forehead_height)

may produce a float value forehead_y_min , while calculating your image size and shape. I detected the place where you got an error.

this solution MAY help you : making the size of the input image to standard-sized may avoid such errors. :) hope may help you :)