Key point detection method and device, computer equipment, medium and program product

Abstract

The invention discloses a key point detection method and device, computer equipment, a medium and a program product, and relates to the technical field of image processing and intelligent traffic. The method comprises the steps that feature extraction is conducted on an original image, a target feature map corresponding to the original image is obtained, the target feature map comprises K category channels and 2M position channels, K is a positive integer, M is an integer larger than 1, the 2M position channels are used for predicting key point positions, the 2M position channels are channels corresponding to unit grids in the target feature map, and M is an integer larger than 1; the unit grids are divided according to M * M; determining a target category channel from the K category channels, wherein a target grid in the target feature map corresponding to the target category channel contains a target key point belonging to a target category; and determining a target key point position of the target key point in the original image based on a target position channel corresponding to the target grid. By adopting the key point detection method, the prediction accuracy of the key point position can be improved.

Description

technical field [0001] The embodiments of the present application relate to the technical fields of image processing and intelligent transportation, and in particular to a key point detection method, device, computer equipment, media and program product. Background technique [0002] Key point detection is one of the common problems in the field of image processing, and it plays an important role in various scenarios such as cloud technology, artificial intelligence, intelligent transportation, and assisted driving. [0003] In related technologies, when deep learning is used to perform key point detection tasks, feature extraction, feature addition, and feature mapping are often performed on the image to be detected to obtain two eigenvalues ​​output by the key point detection model, which represent key points. point location. [0004] Apparently, with the key point detection method in the related art, the key point detection model can only output two eigenvalues ​​to repr...

AI Technical Summary

Problems solved by technology

[0004] Obviously, using the key point detection method in the related art, the key point detection model can only output two eigenvalues ​​to represent the key point position, and the detection accuracy of the key point position is low

Images