Image Classification on PASCAL VOC 2012 (val)
86.5mAPProposed Method
Evaluation results
| Method | Links | |||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Proposed MethodBackbone=VGG-162016.03 | 86.5 | 93 | 89.7 | 91.4 | 89.6 | 69.5 | 90.9 | 81.6 | 92 | 69.3 | 88.9 | 80.2 | — | 90.4 | 90 | 90 | 91.6 | 85.5 | 90.4 | 75.5 | 91.4 | 89.6 | — | |
| TENET2021.03 | 82.3 | 95.6 | 84.3 | 91.1 | 83.1 | 61.3 | 91.4 | 83.2 | 91.6 | 72.8 | 77.4 | — | 65.9 | 91.3 | 84.4 | 89.2 | 96.3 | 57.4 | 83.9 | 67.6 | 94.5 | 83.1 | — | |
| TENETMode=Instance-wise Inhibition2021.03 | 81.3 | 93.1 | 82.7 | 92.6 | 82.9 | 61.1 | 90.9 | 81.8 | 91.6 | 70.6 | 73.7 | — | 63.3 | 91.5 | 85.6 | 88.5 | 96.4 | 56.8 | 85.1 | 61.8 | 93.2 | 82.3 | — | |
| TENETMode=Binary Mask2021.03 | 81.1 | 93.2 | 83.8 | 91.3 | 83.2 | 59.8 | 91.6 | 79.6 | 90.6 | 66.3 | 75.2 | — | 62.1 | 89.7 | 84.7 | 88.4 | 96.3 | 58 | 87 | 65.2 | 93.1 | 82.1 | — | |
| Proposed MethodBackbone=Alexnet2016.03 | 79.8 | 92 | 82.9 | 87.2 | 83.8 | 54.1 | 87.3 | 74.5 | 87 | 56.4 | 76.7 | 71.1 | — | 83.5 | 85.5 | 84.3 | 88.1 | 80.1 | 83.5 | 64.5 | 90.8 | 81.4 | — | |
| GOCNN2021.03 | 79.4 | 96.1 | 81 | 90.8 | 85.3 | 56 | 92.8 | 78.9 | 91.5 | 63.6 | 69.7 | — | 65.1 | 84.8 | 84 | 83.9 | 92.3 | 52 | 83.9 | 64.2 | 93.8 | 78.6 | — | |
| ResNet18Reference=reported in [4]2021.03 | 77.5 | 95.2 | 79.3 | 90.2 | 82.8 | 52.6 | 90.9 | 78.5 | 90.2 | 62.3 | 64.9 | — | 64.5 | 84.2 | 81.1 | 82 | 91.4 | 50 | 78 | 61.1 | 92.7 | 77.5 | — | |
| ResNet18Reference=trained in this paper2021.03 | 77.1 | 94.8 | 83.8 | 91.5 | 79.4 | 56.6 | 88.2 | 78.9 | 90.8 | 64.8 | 61.5 | — | 57.9 | 90.9 | 73.7 | 83.8 | 96 | 51.6 | 77.1 | 58.2 | 89.8 | 77.1 | — | |
| TENET TrainingBackbone=ResNet-18, Protocol=[4], Comparison Baseline=Group Orthogonal Training [4]2021.03 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 2.9 |