This is a person reidentification model for a general scenario. It uses a whole body image as an input and outputs an embedding vector to match a pair of images by the Cosine distance. The model is based on RMNet backbone that was developed for fast inference. A single reidentification head from the 1/16 scale feature map outputs the embedding vector of 256 floats.
| Metric | Value |
|---|---|
| Market-1501 rank@1 accuracy | 0.7791 |
| Market-1501 mAP | 0.6180 |
| Pose coverage | Standing upright, parallel to image plane |
| Support of occluded pedestrians | YES |
| Occlusion coverage | <50% |
| GFlops | 0.028 |
| MParams | 0.280 |
| Source framework | Caffe* |
The cumulative matching curve (CMC) at rank-1 is accuracy denoting the possibility to locate at least one true positive in the top-1 rank. Mean Average Precision (mAP) is the mean across all queries’ Average Precision (AP) and AP is defined as an area under the precision/recall curve.
name: "data" , shape: [1x3x96x48] - An input image in the format [BxCxHxW], where:
The expected color order is BGR.
[*] Other names and brands may be claimed as the property of others.