Quantitative evaluation method of lower limb multiparameter fusion based on BP neural network

Le résumé fourni par la source

Aiming at the problems of subjectivity and insensitivity of assessment results in the traditional rehabilitation assessment process of the lower limbs of hemiplegic patients, this manuscript proposes a multi-parameter fusion objective rehabilitation assessment method for the lower limbs based on the back-propagation neural network (BP). The subject's lower limb motion data were collected by a motion capture system, and five indexes such as gait symmetry and joint range of motion were selected as inputs, and the assessment scores as outputs, to establish a BP neural network based multiparameter fusion rehabilitation evaluation model for lower extremities. The BP neural network model was optimized by Newton-Raphson-based optimizer (NRBO) to further improve the accuracy of the evaluation model. To compare the improved neural network (NRBO-BP) with the BP neural network, Convolutional Neural Network (CNN), and Long Short-Term Memory Network (LSTM), and validate the effectiveness of the optimized BP neural network model for lower limb functional rehabilitation assessment.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.