Wi-Fi dataset for CGS-CMGR

To evaluate the performance of CGS-CMGR, we construct a dedicated gesture recognition dataset containing 7,600 single-user and 1625 multi-user gesture samples. The dataset is collected from five volunteers across two distinct environments, covering nine different positions and two LoS conditions. It comprises five gesture categories: Push, Circle, Triangle, M, and Square. A detailed description of the dataset is provided in the table below. NameEnvironmentTotal samplesGestureNumber of LocationsNumber of UsersNumber of LoS conditionsNumber of InstancesSingle-user dataset 1Basement6000(a) Push, (b) Circle,(c) Triangle, (d) M, (e) square65220Single-user dataset 2Conference room160042220Dual-user datasetBasement500Any combination of two gestures (25 combinations)Fixed LocationsTwo fixed usersNLoS10Tri-user datasetBasement1125Any combination of three gestures (125 combinations)Fixed LocationsThree fixed usersNLoS3