"MORSE-Face Dataset: A Multi-Distance, Occlusion-Rich, Real and Spoof Evaluation Face Dataset"

"Face presentation attacks (also known as spoofing attacks) remain a critical challenge for the reliability and security of face recognition systems, particularly under unconstrained conditions involving occlusion, long capture distances, and heterogeneous acquisition devices. To support systematic research in this area, we introduce MORSE-Face Dataset, a large-scale, multi-distance, occlusion-rich face dataset designed for real and spoof face analysis. The dataset was collected at Clarkson University between January 2025 and February 2026 and consists of data from 250 volunteer subjects captured across 283 recording sessions in both indoor and outdoor environments. MORSE-Face includes synchronized image and video data acquired using a professional DSLR camera and a modern smartphone, with capture distances ranging from 1 m to 100 m. The dataset covers a comprehensive set of face presentation conditions, including normal live faces, multiple categories of occlusions (hair alterations, facial obstructions, and eyewear), two types of two-dimensional spoof artifacts (printed photographs and color cloth masks), and two types of three-dimensional spoof masks (3D printed mask and wearable silicone mask), both with and without additional props. Each subject contributes data under multiple controlled variations, enabling fine-grained analysis of spoofing behavior across distance, environment, device, and occlusion factors. In addition to biometric data, MORSE-Face dataset provides rich demographic metadata, including age, gender, and race, enabling research beyond presentation attack detection, such as demographic bias analysis and age or gender prediction. With over 15,000 images and 15,000 videos totaling approximately 1.3 TB, MORSE-Face represents one of the most comprehensive datasets for studying real-world face presentation attacks. The dataset is intended to support academic research in presentation attack detection, occlusion-robust face recognition, long-distance face analysis, and cross-device biometric evaluation."