US Patent 11403730 Method and apparatus for processing image

Techniques for processing an image are described herein. The disclosed techniques include a target human body image is acquired, and a skeleton key point data set is determined from the target human body image. Based on the skeleton key point data set, a target region in the target human body image is determined, and for each of pixel points comprised in the target region, performing coordinate transformation is performed on the pixel point, to generate a transformed coordinate of the pixel point. With the embodiment, different coordinate transformations may be performed on pixel points of different regions in the target human body image, thereby improving flexibility of performing the coordinate transformation on the pixel points.