"Real-Time Gesture Classification using ToF Sensor and Embedded 1D-CNN"

"This public dataset was created as part of the Autonomous Computing course and is intended for training ML (Machine Learning) models capable of recognizing hand gestures in real time using a ToF (Time-of-Flight) sensor. The goal is to enable touchless human-machine interaction without relying on traditional RGB cameras. The dataset contains a total of 8,400 samples collected with an STMicroelectronics VL53L5CX sensor mounted on a B-U585I-IOT02A development board. The sensor captures depth information through an 8x8 grid, generating 64 distance measurements per frame while preserving user privacy. Data collection was performed in a controlled environment with the sensor fixed in position while different hand gestures were executed in front of it. The dataset includes four gesture classes: Hold, Tap, Approaching, and Swipe, with 2,100 balanced samples per gesture. Data was recorded at a sampling frequency of 32 Hz, and each entry includes the depth matrix values, relative time, and a precise timestamp."