VT-KFER dataset is a RGBD+time facial expression recognition dataset collected using the Kinect 1.0 sensor in both scripted (acted) and unscripted (spontaneous) scenarios. This fully annotated dataset includes seven expressions (happiness, sadness, surprise, disgust, fear, anger, and neutral) for 32 subjects (males and females) aged from 10 to 30 and with different skin tones. Both human and machine evaluation were conducted. Each scripted expression was ranked quantitatively by two persons.Related publication:Sherin Aly, Andrea Trubanova, Lynn Abbott, Susan White, and Amira Youssef, "VT-KFER: A Kinect-based RGBD+Time Dataset for Spontaneous andNon-Spontaneous Facial Expression Recognition", International Conference of Biometrics, 2015.[PDF]