RuntimeWarning: в double_scalars обнаружено недопустимое значение ret = ret.dtype.type (ret / rcount) оценка среднего тона кожи (BGR) nan nan nan []

Question

сегментация переднего и заднего плана (otsu) и преобразование изображения в цветовые пространства HSV и YCrCb и обнаружение потенциальных пикселей

blue = []
green = []
red = []
height, width, channels = img_face_only.shape
print('width:  ', width)
print('height: ', height)
print('channel:', channels)

for i in range (height):
   for j in range (width):
       if((img_HSV.item(i, j, 0) <= 170) and (140 <= img_YCrCb.item(i, j, 1) <= 170) and 
                                      (90 <=img_YCrCb.item(i, j, 2) <= 120)):

           blue.append(img_face_only[i, j].item(0))

           green.append(img_face_only[i, j].item(1))

           red.append(img_face_only[i, j].item(2))


       else:
           img_face_only[i, j] = [0, 0, 0]

display_image(img_face_only, "final segmentation")

# determine mean skin tone estimate
skin_tone_estimate_BGR = [np.mean(blue), np.mean(green), np.mean(red)]
print ("mean skin tone estimate (BGR)", skin_tone_estimate_BGR[0], skin_tone_estimate_BGR[1], skin_tone_estimate_BGR[2], "]")```


> the result is always a [nan nan nan] which is wrong, I need to return the bgr values.