Rehabilitation method and system based on multi-sensor action recognition

Abstract

The invention relates to a rehabilitation method and system based on multi-sensor action recognition. The method specifically comprises the following steps: a sample data obtaining step: obtaining acceleration sample data of rehabilitation actions and adding the acceleration sample data into a database; a local feature obtaining step: according to the acceleration sample data in the database, adopting a CART decision tree to perform classification training to obtain a plurality of local features; an action identification step: according to synchronously acquired acceleration signal data and surface electromyogram signal data, realizing action identification according to a plurality of the local features and single-dimensional features of the surface electromyogram signal data; and a compensation step: implementing myoelectricity compensation according to the identified actions and the surface myoelectricity signal data. A multi-sensor data fusion algorithm used for human body intention recognition is adopted, all local features are re-assembled into complete one-dimensional features through a matrix by using a full connection layer, the one-dimensional features are mapped into probabilities by using a soft-max layer, and classification is completed.

Description

technical field [0001] The invention relates to a rehabilitation method and system based on multi-sensor action recognition. Background technique [0002] In recent years, the number of patients with lower extremity motor dysfunction caused by central nervous system diseases such as spinal injuries and strokes has increased sharply, seriously endangering human health. With the development of society and the improvement of people's medical care and living standards, the health of the disabled has attracted the attention of the whole society. Weight-loss walking training is one of the important means of walking rehabilitation for patients with such diseases, and a large number of clinical studies have confirmed its effectiveness. Traditional rehabilitation treatment methods mainly rely on nurses to assist patients in rehabilitation training. The effect of rehabilitation training depends on the technical level and love of nurses. At the same time, the number of nurses is serio...

Images