CV

Images are captured by professional photographers. Composition types include rule-of-thirds, horizontal, diagonal, triangular, and central composition. All images are evaluated and annotated by personnel with high aesthetic standards. Each image meets at least one composition type and at most three composition types.

The images of human figures in this dataset have been transformed into four different styles based on the original human images, including 3D cartoon, comic, watercolor painting, and sketch. There are four skin colors included: dark, fair, brown, and light. Each style-generated image contains all four skin colors.

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.

This dataset collects single-person photos/combined photos. The subjects are grouped into either single-person or 2-4 person combinations for shooting. The shooting scenes cover both indoor and outdoor environments, with various types of lighting. The subjects include various skin tones, face shapes, and eye shapes.

This dataset collects facial images in both RGB and IR formats. The ages included are children (7-16 years old) and the elderly (over 60 years old) as well as adults. The lighting conditions captured range from normal light, strong light to dim light. The facial postures of the collectors include normal facial state and five specific postures (such as frontal face, 30 degrees to the left, etc.).

This dataset covers more than 10 types of scenarios, including residential settings, offices, retail stores, restaurants, manufacturing facilities, warehouses, agricultural environments, and entertainment venues. It includes over 200 tasks and provides rich first-person action and human–environment interaction data. The videos have a resolution of 1920 × 1080 and a frame rate of 60 fps. The dataset can be used for embodied AI model training, egocentric behavior understanding, environmental perception and interaction, and other related applications.