[FIXED] What's the definition of image size in OpenCV?

Issue

What's the size

import cv2

img = cv2.imread('./imgs/img.png', cv2.IMREAD_COLOR)

print(img.shape[0], img.shape[1])  # 182 277
print(img.size)  # 151242

I thought img.size = img.width * img.width, but obviously wrong, check this example.

so, what's the definition of it?

Solution

It's all about what language you're working in.

The Python interface to OpenCV uses numpy, and numpy arrays have different methods from cv::Mat (C++). size means different things then.

• Python: OpenCV uses numpy arrays. .size tells you the number of elements in the array. That is the product of all values in .shape.

• C++: OpenCV uses its own cv::Mat. .size there is of type MatSize. It is roughly equivalent to .shape with numpy. Documentation indicates that it can also tell you the number of channels. It has an operator[] defined.

Answered By - Christoph Rackwitz