Zero-shot Object Detection on ZSFooD 1.0 (test)
0.87Recall@100SeeDS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SeeDSEvaluation Protocol=GZSD - Seen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.87 | 82.8 | |
| RRFSEvaluation Protocol=GZSD - Seen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.866 | 82.7 | |
| CZSDEvaluation Protocol=GZSD - Seen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.861 | 81.2 | |
| SUEvaluation Protocol=GZSD - Seen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.823 | 79.1 | |
| SeeDSEvaluation Protocol=GZSD - Harmonic Mean, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.633 | 6.7 | |
| RRFSEvaluation Protocol=GZSD - Harmonic Mean, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.614 | 5.2 | |
| ConSEEvaluation Protocol=GZSD - Seen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.58 | 54.3 | |
| SUEvaluation Protocol=GZSD - Harmonic Mean, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.574 | 4.5 | |
| BLCEvaluation Protocol=GZSD - Seen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.553 | 51.1 | |
| CZSDEvaluation Protocol=GZSD - Harmonic Mean, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.548 | 4.1 | |
| SeeDSEvaluation Protocol=ZSD, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.529 | 5.9 | |
| SeeDSEvaluation Protocol=GZSD - Unseen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.498 | 3.5 | |
| RRFSEvaluation Protocol=ZSD, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.488 | 4.3 | |
| CZSDEvaluation Protocol=ZSD, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.48 | 4 | |
| RRFSEvaluation Protocol=GZSD - Unseen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.476 | 2.7 | |
| BLCEvaluation Protocol=GZSD - Harmonic Mean, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.468 | 1.8 | |
| ConSEEvaluation Protocol=GZSD - Harmonic Mean, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.464 | 1.4 | |
| SUEvaluation Protocol=ZSD, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.453 | 3.9 | |
| CZSDEvaluation Protocol=GZSD - Unseen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.448 | 2.1 | |
| SUEvaluation Protocol=GZSD - Unseen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.441 | 2.3 | |
| BLCEvaluation Protocol=ZSD, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.412 | 1.1 | |
| BLCEvaluation Protocol=GZSD - Unseen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.405 | 0.9 | |
| ConSEEvaluation Protocol=ZSD, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.397 | 0.8 | |
| ConSEEvaluation Protocol=GZSD - Unseen, Backbone=ResNet-101, Word Embedding=CLIP2023.10 | 0.381 | 0.7 |