MSRC-12 stands for The Microsoft Research Cambridge-12 Kinect gesture dataset, which is a dataset used to study gesture recognition. The dataset consists of 594 motion sequences and 719,359 frames of related actions that can be recognized by the system, totaling about 6 hours and 40 minutes. A total of 30 volunteers participated in the collection of the dataset, and each person was asked to make 12 gestures, and a total of 6,244 gesture instances were collected.