Thank you for your contribution of the code. I completed the detection of 4 key points according to your code, but the second coordinate and the third coordinate overlap. May I ask why, How to solve this problem
This model was trained on predicting joints in humans. Therefore, I wouldn't expect it to be best for generic detection of keypoints e.g. of a piece of paper. I'd recommend using a generic detection method such as the YOLO line of work.
Thank you for your contribution of the code. I completed the detection of 4 key points according to your code, but the second coordinate and the third coordinate overlap. May I ask why, How to solve this problem