The DukeMTMC-VideoReID (Duke Multi-Tracking Multi-Camera Video-based ReIDentification) dataset is a subset of the DukeMTMC for video-based person re-ID. The dataset is created from high-resolution videos from 8 different cameras. It is one of the largest pedestrian video datasets wherein images are cropped by hand-drawn bounding boxes. The dataset consists 4832 tracklets of 1812 identities in total, and each tracklet has 168 frames on average.
NOTE: This dataset has been retracted.
Source: Exploit the Unknown Gradually: One-Shot Video-Based Person Re-Identification by Stepwise Learning
Variants: DukeMTMC-VideoReID
This dataset is used in 1 benchmark:
| Task | Model | Paper | Date |
|---|---|---|---|
| Person Re-Identification | STRF | Spatio-Temporal Representation Factorization for Video-based … | 2021-07-25 |
Recent papers with results on this dataset: