A dataset of exocentric images capturing hand gestures in multiplayer games
Year: 2025
Type: data-paper
Abstract: This dataset presents 7221 exocentric (third-person) frame images of hand gestures recorded during multiplayer tabletop games including Ludo, Poker, and Snakes and Ladders. The recordings were made using a Google Pixel 6a smartphone at 30 frames per second (FPS), and frame images were extracted at a fixed interval of one frame every four seconds. Each extracted image was resized to 512 × 512 pixels. The dataset captures two-, three-, and four-person gameplay sessions under four distinct lighting conditions (natural, white, yellow, and dim), reflecting realistic and varied visual environments. The frames include natural gesture dynamics with motion blur, occlusions, and object manipulations, offering authentic interaction data. This dataset supports research in computer vision, human-computer interaction, and Extended Reality (XR), enabling studies in gesture recognition, hand-object interaction, and occlusion handling. It can also be applied in supervised and unsupervised learning tasks, evaluation of gesture-tracking algorithms, and XR benchmarking for real-world gaming interactions. By providing authentic multi-person interaction footage from consistent top-down viewpoints, this dataset facilitates advancements of gesture recognition systems designed for collaborative environments and interactive gaming applications.
Source: Data in Brief
Institution: American International University-Bangladesh
Language: English
Subfield: Human-Computer Interaction
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,220
OpenAlex ID: W4416106121