This invention relates to a method for visual skeleton
rehabilitation movement assessment and
occlusion completion. It acquires a
video sequence, extracts a preset set of key points and their confidence levels frame by frame, and generates a
human body mask. Based on the confidence level, the difference in length between adjacent bone segments, the difference in key point displacement, and whether the key point is located within the
human body mask, key points are determined to be reliable, questionable, or unreliable. Key points that are considered questionable or unreliable in a consecutive preset positive integer N frames are identified as
occlusion candidates, and
occlusion regions are formed according to the skeleton's
connectivity. Based on the distance between the occlusion region and the
image boundary and the
connectivity of the
human body mask, the occlusion cause is classified as out-of-frame, self-occlusion, or external object occlusion. While keeping reliable key points unchanged, completion candidates are generated according to the cause, and candidates that do not meet the bone
segment length deviation threshold, displacement jump threshold, or are located outside the human body mask are eliminated. The completed preset set of key points of the
human skeleton is output. This improves the accuracy of occlusion judgment and completion, and enhances the reliability of
rehabilitation movement assessment.