iPer Portrait Synthesis Dataset

iPer, also known as Impersonator, is a video dataset with various styles of clothing in the video, which is used to evaluate human motion imitation, appearance transfer, and novel view synthesis. A total of thirty volunteers participated in the data collection, with different body shapes, heights, and genders. Each volunteer wore different clothing, recorded a video of pose A and a video of random movements, totaling 103 pieces of clothing. The dataset has a total of 241,564 frames, including 206 video sequences.