Skeleton Based Hand Missing Score Evaluation System

Abstract

In this paper, a skeleton-based method is proposed for evaluating hand defects. Specifically, we take use of the existing hand skeleton extraction method and correct the extracted skeleton points, making them applicable for a hand with defects. Then, we combine the method above with the existing human hand defect evaluation system to automatically evaluate the concrete situation of the hand. The proposed method is tested on a dataset constructed by ourselves, which includes various cases where fingers and palm are missing. Experimental results illustrate that our method achieves excellent performance and can be applied to scenarios where the hand is partially missing with high accuracy. Our method can be helpful to forensic agencies in detecting and evaluating hand mutilations.

Lab Members

Patrick Chan

Associate Professor, Vice Dean

Patrick Chan works on machine learning, deep learning, image processing, adversarial learning, and secure machine learning.