The Application of Multimedia and Deep Learning in the Integration of Professional and Innovative Education in Colleges

Abstract

In this paper, the authors study the assessment of professional practice courses in the integrated education of specialization and innovation, and use the multimedia and deep learning technology to complete the action recognition of students in practice courses. Firstly, the skeletal features are extracted from multimedia video data by Openpose algorithm, which is used for subsequent classification while ensuring privacy; then the LSTM method is used to recognize typical motions in student practice, and the average recognition result exceeds 89%; finally, practical application tests are conducted for laboratory and office scenes, and the results illustrate that the proposed framework performs well in the tests with recognition rate exceeding 80%. The algorithm framework provides a new idea for the curriculum setting and evaluation method of professional practice education, and gives data guarantee for their integration and innovation education.