[Python] How to use opencv to detect the number of items in a picture?

I want to get return value "1" for image1, and "2" for image2, (because there is a hand in image2 compared to image1), or at least to get a return value from image2 bigger than the value of image1.

I have tried the following code:

import cv2
import sys

filename = sys.argv[1]

# 1. Load and Preprocess Image
img = cv2.imread(filename)


# Convert to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
gray = cv2.GaussianBlur(gray, (15, 15), 0)

cv2.imwrite('gray.jpeg', gray)

edged = cv2.Canny(gray, 5, 100)
cv2.imwrite('edged.jpeg', edged)


# Find the contours of the object
contours, hierarchy = cv2.findContours(edged, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

print(len(contours))


but it gives "135" for image1, and "62" for image2. The value of image2 is less than that of image1. Some internet posts say YOLO could detect objects, but I tried and it seemed YOLO can only detect predefined classes. Thanks a lot for any response.

Image 1

Image2

Continue reading...