Parkinson hand motion quantitative analysis method and system based on depth image

Abstract

The invention discloses a Parkinson's hand motion quantitative analysis method and system based on a depth image, and belongs to the field of computer vision and machine learning. The method comprisesthe steps that a detected person makes a hand action according to the requirements of a Parkinson's disease evaluation table, multiple frames of depth images of the detected person are acquired in the period, and 3D coordinates of the hand centroid in each frame of depth image are recognized; according to the 3D coordinates of the hand centroid, hand point cloud and noise point cloud segmentationis carried out on each frame of depth image; predicting 3D coordinates of each hand joint point in each frame of depth image based on the hand point cloud in the single frame of depth image; integrally optimizing the 3D coordinates of all the hand joint points of all the frames according to the time sequence information between the continuous frames and the priori knowledge of the hand postures;according to the optimized articulation point 3D coordinates, hand motion features are extracted; and utilizing the trained XGBoost classifier to classify the extracted hand motion features, and giving a corresponding scoring result.

Description

technical field [0001] The invention belongs to the field of computer vision and machine learning, and more specifically relates to a method and system for quantitative analysis of Parkinson's hand movements based on depth images. Background technique [0002] The onset symptoms of early Parkinson's disease are not obvious, and the traditional diagnosis method requires a series of judgments. Among them, a key link is to judge based on the completion status of a series of specified actions. One of the traditional diagnostic methods is to evaluate patients one by one through the Parkinson's Rating Scale UPDRS. The doctor guides the patient to complete the actions according to the Parkinson Rating Scale, and then scores item by item according to the patient's completion, which takes about 30 minutes or even longer. During the period, you may suffer mental damage from self-doubt and doctor's words due to physical coordination, tension and other factors. This method of diagnos...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problems that the prior art cannot protect the privacy of patients and the accuracy of motion assessment is poor, the present invention provides a quantitative analysis method and system for Parkinson's hand motion based on depth images, the purpose of which is to effectively protect patients based on the analysis of depth images Privacy; Quantitative analysis of hand motion based on the three-dimensional motion of joint points in physical space, and based on the timing information between consecutive frames and the prior knowledge of hand posture, the 3D coordinates of all hand joint points in all frames Overall optimization to improve evaluation accuracy

Images