The application provides an electromyogram
muscle strength prediction method suitable for
hand function rehabilitation training of children with
cerebral palsy, and specifically comprises the following steps: constructing a
muscle strength prediction source domain
network model, and using an electromyogram
muscle strength
database to
train and test the model; proposing a gesture set and a synchronous
acquisition scheme of electromyogram signals and
muscle strength signals of each gesture according to the
rehabilitation training requirements of children with
cerebral palsy; collecting electromyogram and
muscle strength data generated when healthy children perform each gesture action, optimizing the
muscle strength prediction source domain
network model by using a transfer learning method, and obtaining a specific muscle strength prediction model for each gesture; collecting electromyogram signals and muscle strength signals of children with
cerebral palsy during
rehabilitation training of different gestures, realizing muscle strength prediction by using the specific muscle strength prediction model, and evaluating the completion effect of rehabilitation gestures of children with cerebral
palsy according to the error between the predicted value and the actual value. The method is conducive to formulating a specific rehabilitation scheme and evaluating the rehabilitation
treatment effect by clinicians, and better promotes the rehabilitation treatment of children with cerebral
palsy.