Multi-Scenario Human Motion Recognition Framework Combining Skeleton Key Points Extraction and Classification Modeling

The recognition of human body movements is an important basis for analyzing human intentions. Enabling computers to observe and comprehend human body motions is extremely important. Currently, human motion recognition technology is extensively utilized in applications such as human-computer interface, medical rehabilitation training, and fitness activity correction. It is challenging to accurately identify human movement in different scenarios. The skeleton key point extraction and classification modeling are used in this study to construct a human motion recognition system. The network is trained by extracting data from human key skeletal points. At the same time, we carried out comparative experiments with different classifiers, as well as evaluated the generalization in different scenarios and on different testers.