Open hector6298 opened 10 months ago
Objective: Create a small training set from source pool that can train a model with high accuracy on target data.
Look at the size of the blob (it represent the number of identities), it is comparable to other datasets like Market and Duke, but achieves higher accuracy.
It can be used with our existing architecture, or their baseline (Google Inception)
Publications and repositories
Search and Pruning (SnP) Framework for Training Set Search
Parsing-based viewaware embedding network for vehicle ReID
Torchreid