ZJU-MV is the dynamic multi-view companion dataset to ZJU-OLAT.
The release is organized as one tar archive per synchronized sequence. Each
archive contains per-view H.264 MP4 videos at 25 fps together with image/video
size reports and per-view compression metadata.
Additional… See the full description on the dataset page:
https://huggingface.co/datasets/STLAND/ZJU_MV.