The release version of SBeA.
Linked paper: Multi-animal 3D social pose estimation, identification and behaviour embedding with a few-shot learning framework
For the single animal usage, please jump to Behavior Atlas, and the related paper is A hierarchical 3D-motion learning framework for animal spontaneous behavior mapping

Notes:
Blue color: animal 1
Red color: animal 2
Masks: the segmentation of them
Bounding boxes: generated by the maximum external boxes of masks
Left: Video capture with multi-view cameras
Center: Cropped identities
Right: 3D poses with identities of multiple animals

Notes:
Left three frames: the free-social interactions of three group of mice
WT: Wild type mouse
KO: Shank3B knockout mouse
Right: social behavior atlas with distance map
Circle: current social behavior module of left three groups
Trace: previous social behavior modules last 100 frames
(1) Shank3B and WT mice social interaction videos
Baidu Netdisk:
Link: https://pan.baidu.com/s/1H1a4oXHWQV0SThgId_4yBA?pwd=2iyk
Extraction code: 2iyk
figShare:
social videos: https://figshare.com/articles/dataset/shank3b_wt_social_videos/30338929
id videos: https://figshare.com/articles/dataset/shank3b_wt_id_videos/30338953
(2) SBM-VIS-VIStR dataset with animal IDs
figShare: https://figshare.com/articles/dataset/SBM-VIS-VIStR-20251103/30514163

