RGB-D Facial Dataset of Young Adults with 9-Directional Gaze and Expression Variations for 3D Biometrics and Computer Vision

This work presents a specialized RGB-D facial dataset developed to provide a robust benchmark for 3D biometric analysis and computer vision research. The collection comprises synchronized color and depth information from 57 subjects within the 18–30 age demographic. The core of the dataset is structured around a comprehensive 3×3 orientation matrix, capturing nine distinct gaze directions that combine horizontal and vertical facial poses. To address physiological and behavioral variability, each pose was recorded across three emotional states: neutral, smile, and scowl. Furthermore, the dataset implements a multi-frame burst acquisition strategy (9 samples per state at 0.1s intervals), resulting in a dense repository of facial geometry and texture. By providing raw per-pixel distance measurements in an accessible plain-text format, this database facilitates the development of expression-invariant and pose-robust recognition algorithms. The complete dataset, including structured metadata and a standardized visualization pipeline, is openly availble to the scientific community to support the advancement of 3D facial modeling and identification.