This dataset contains a collection of adult and teenage facial expression images. The indoor scenes feature plain colors or other patterns as the background fabric, and three cameras are fixed to capture the images, with different angles of elevation. Each collector performs 14 facial expressions. The collector's race is of Asian descent, with the age range from 13 to 80 years old, and the gender ratio is roughly balanced. The shooting distance is head and shoulder shots or half-body shots.
This dataset is a collection of indoor expression images for people of various skin tones. The main skin tones are: Light, white, and Dark. The indoor scenes have a solid color background cloth, and the cameras are fixed for shooting, with different angles. Each collector completes 14 expression actions. The collection equipment includes two types of cameras: the first folder contains 720P cameras, and the second folder contains 1080P cameras.
Indoor children's expression image dataset. The indoor scene has a solid color or other patterned background cloth. Three cameras are fixed for shooting, with different angles. Each collector completes 14 expression actions. The collector's race is of Asian descent, the collected age range is 6 to 12 years old, and the gender ratio is balanced. The shooting distance is head and shoulder shots or half-body shots.
This dataset is a collection of images captured through indoor eye-tracking. The indoor scenes feature solid colors or other patterns as the background fabric, and three cameras are fixedly positioned for shooting, with different angles of elevation. The subjects include Chinese adults, Chinese children. The gender ratio is roughly balanced.
This dataset collects real-person data and five types of attack data for each ID. The attack methods include photo, image mask, ID headshot, mask (with real human features exposed), and video dynamic attack, etc. The image data includes multiple formats such as RGB, IR (infrared), and Depth (depth). The collection equipment is Hua Jie, Intel, and Diwei Tai (the video deliverables from these three types of equipment are image sequences, not videos). The image sequence (which can be understood as a series of consecutive frames).
