This dataset serves as a comprehensive resource for developing and testing driver monitoring systems, specifically designed to enhance in-vehicle safety through behavioral analysis.
The product library size is approximately 45TB, containing over 23,950,000 files.
Gender
The data comprises 70% males and 30% females.
Collection environment
The images capture for action and gaze includes 30 gaze positions inside the car, with 11 fixed camera positions and 19 mobile camera positions. When the gaze is directed towards a fixed camera position, 25 fixed camera images are captured simultaneously; when the gaze is directed towards a mobile camera position, 25 fixed camera images and 1 mobile camera image are captured simultaneously.
Collectors
700+ People
Ethnic Diversity
Categorized into Black (encompassing black, brown, olive skin tones) and White (including fair, medium, very fair complexions)
Global Representation
39+ nationalities spanning Switzerland, Colombia, Peru, Ghana, Brazil, Latvia, Samoa, South Africa, and Latin American regions
Product Features: Captured in a conference setting, with participants maintaining a neutral facial expression throughout, slowly walking around the room, without side glances or looking up/down, and with faces unobstructed. Each participant records 1–2 sets of videos (standing/sitting) using four cameras simultaneously (Logitech Rally, Aver CAM 550, Yealink UVC86, Poly E60). Additionally, two sets of photos are collected: one taken on the day using a computer, and one personal photo taken within the past two years using a phone.
Ethnicities Collected: Black, White, Asian (non-Chinese), Brown.
Age Range: All age groups, with balanced gender ratio.
Video/Image Specifications: Video resolution 1080P or higher, photo resolution 720P; each video is approximately 1 minute long.
Android Front-Facing Multi-Skin-Tone Face Collection Corpus
Product Features: Captured using the front-facing camera of Oppo series smartphones released after 2023. Models maintain eye contact with the camera, with a shooting distance of 20–50 cm (selfie unlock distance). Includes six lighting conditions (normal light, side light, backlight, front light, low light, warm light). For each participant: at least 15 photos for Black, White, and Asian (Chinese) individuals; at least 25 photos for Brown individuals. The dataset includes 25 pairs/groups of twins or look-alike individuals.
Ethnicities Collected: Black, White, Asian (Chinese), Brown.
Age Range: All adults.
Image Specifications: 1080P resolution or higher.
Product Features: AIGC-generated portrait data covering four styles—3D cartoon, comic, watercolor, and sketch—with each style including four ethnicities. The original portrait images are shared across styles, with about 10 images generated per person for each style.
Ethnicities Collected: Black, White, Asian (non-Chinese), and Brown.
Age Range: All adults, with a balanced gender ratio.
Image Specifications: Resolution of 1080P or higher.
Product Features: This corpus includes data from 23 categories such as cuisine, landscapes, architecture, cities, countryside, health, sports, medical, automobiles, backgrounds, finance, education, oil paintings, illustrations, watercolors, travel, fashion, romance, animals, plants, space, and technology.
