Zero-shot recognition on aPascal & aYahoo (aP&Y)
43.1Top-1 AccuracyLisGAN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LisGANLearning Setting=Zero-shot2019.04 | 43.1 | — | |
| GAZSLLearning Setting=Zero-shot2019.04 | 41.1 | — | |
| f-CLSWGANLearning Setting=Zero-shot2019.04 | 40.5 | — | |
| DeViSELearning Setting=Zero-shot2019.04 | 39.8 | — | |
| ALELearning Setting=Zero-shot2019.04 | 39.7 | — | |
| ESZSLLearning Setting=Zero-shot2019.04 | 38.3 | — | |
| DEMLearning Setting=Zero-shot2019.04 | 35 | — | |
| SSELearning Setting=Zero-shot2019.04 | 34 | — | |
| DAPLearning Setting=Zero-shot2019.04 | 33.8 | — | |
| SJELearning Setting=Zero-shot2019.04 | 32.9 | — | |
| CONSELearning Setting=Zero-shot2019.04 | 26.9 | — | |
| SYNCLearning Setting=Zero-shot2019.04 | 23.9 | — | |
| SAELearning Setting=Zero-shot2019.04 | 8.3 | — | |
| Farhadi et al. [8]Feature=Non-CNN2015.09 | — | 0.325 | |
| Jayaraman and Grauman [14]Feature=Non-CNN2015.09 | — | 0.2602 | |
| Lampert et al. [16]Feature=Non-CNN2015.09 | — | 0.191 | |
| Lampert et al. [16]Feature=vgg-verydeep-192015.09 | — | 0.3816 | |
| Mahajan et al. [18]Feature=Non-CNN2015.09 | — | 0.3793 | |
| Romera-Paredes and Torr [30]Feature=Non-CNN2015.09 | — | 0.2727 | |
| Romera-Paredes and Torr [30]Feature=vgg-verydeep-192015.09 | — | 0.2422 | |
| SSE-INTFeature=vgg-verydeep-192015.09 | — | 0.4415 | |
| SSE-ReLUFeature=vgg-verydeep-192015.09 | — | 0.4623 | |
| Wang and Ji [39]Feature=Non-CNN2015.09 | — | 0.4505 |