Application of multi feature medical image fusion in biomechanics experimental teaching of sports rehabilitation

With the increasing demand of society for sports rehabilitation, many colleges and universities have set up the sports rehabilitation biomechanics specialty, which has been listed as a professional basic course in some colleges and universities. However, there are still many problems in the experimental teaching of this course, such as the failure to recognize the human movements in medical images in a timely manner, which makes it impossible to judge whether the movements of motion rehabilitation are standard. However, multi feature medical image fusion can become a new method, which has been widely used in the application and research of medical images, and its position and role in teaching have been further strengthened. Multi feature medical image fusion can not only eliminate redundant information, but also maintain the original image information, and achieve better fusion results. Under the background of multi feature medical image fusion, this paper proposed a support vector machines (SVM) method based on mixed kernel function. This method not only makes the motion recognition of traditional Chinese medical images of motion rehabilitation biomechanics more accurate, but also faster. The experimental results showed that the recognition accuracy and recognition rate of the kernel function were 60.2% and 53.7% when the number of samples was 800. The recognition accuracy and recognition rate of SVM were 65.5% and 53.8%. The recognition accuracy and recognition rate of SVM based on mixed kernel function were 89.9% and 86.7%. This further proved that SVM based on mixed kernel function was superior to the other two methods in recognition accuracy and recognition rate, which proved the superiority of this method in rehabilitation motion recognition.