Hand Recognition Dataset for Machine Vision Researchers (YOLOv8 Format)

This hand recognition dataset comprises a comprehensive collection of hand images from 65 individuals, including both left and right hands, annotated with YOLOv8 formatting.

The dataset encompasses 17 distinct classes, denoted as L-L1 to L-L9 for the left hand and R-R1 to R-R8 for the right hand. These classes capture various hand gestures and poses.

These images were captured using a standard mobile phone camera, offering a diverse set of images with varying angles and backgrounds. In total, the dataset comprises 405 high-quality images, with 222 representing left hands and 183 representing right hands. The left hand classes are distributed as follows: L-L1 (62 images), L-L2 (56 images), L-L3 (44 images), L-L4 (29 images), L-L5 (14 images), L-L6 (8 images), L-L7 (4 images), L-L8 (2 images), and L-L9 (3 images). Similarly, the right hand classes are distributed as R-R1 (53 images), R-R2 (48 images), R-R3 (38 images), R-R4 (24 images), R-R5 (14 images), R-R6 (4 images), R-R7 (1 image), and R-R8 (1 image).

We welcome the machine vision research community to utilise and build upon this dataset to advance the field of hand recognition and its applications.