Vision-based estimation of MDS-UPDRS scores for quantifying Parkinson's disease tremor severity

Weiping Liu, Xiaozhen Lin, Xinghong Chen, Qing Wang, Xiumei Wang, Bin Yang, Naiqing Cai, Rong Chen, Guannan Chen, Yu Lin

Abstract: Highlights•We propose a video-based method for estimating PD tremor severity scores using tremor videos and classify them into clinical scores.•EVM was innovatively used for the pre-processing of tremor video to magnify subtle tremors in the video.•Our proposed GSM has a significant superiority in the identification of tremors compared to other methods.•Our method (EVM+GTSN) outperforms other state-of-the-art models in making score prediction on a fairly large dataset.