Table 4.
Comparison of the proposed ER method with state-of-the-art methods for 10 + 10 classes scenario. Best results are written in bold number.
| Methods | Aero | Bike | Bird | Boat | Bottle | Bus | Car | Cat | Chair | Cow | Din. Table | Dog | Horse | M Bike | Person | Plant | Sheep | Sofa | Train | TV | mAP at 0.5 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| All data | 76.5 | 84.4 | 68.9 | 59.3 | 55.6 | 82.1 | 85.9 | 79.9 | 55.4 | 73.7 | 67.8 | 79.9 | 85.2 | 84.4 | 83.1 | 46.1 | 71 | 70.3 | 82.1 | 71.7 | 73.2 |
| KD [1] | 10.8 | 10.2 | 36.1 | 9.6 | 6.3 | 5.4 | 14.3 | 41.7 | 3.8 | 19.4 | 42.5 | 53.5 | 76.5 | 66.8 | 57.9 | 27.8 | 60.4 | 50.4 | 72.2 | 46 | 35.5 |
| DMC [17] | 56.6 | 46.5 | 48.7 | 25.7 | 45.5 | 62.9 | 71.5 | 65.7 | 36.8 | 52.4 | 0.6 | 60.1 | 71.2 | 58.2 | 73 | 40.7 | 58.9 | 12.6 | 64 | 66.6 | 50.9 |
| GEM [15] | 3.1 | 3.4 | 3.6 | 6.1 | 19.7 | 6.1 | 58.6 | 9.5 | 55.7 | 14.2 | 0.7 | 7.6 | 4.2 | 2.9 | 23.5 | 0.2 | 1.2 | 2.6 | 3.4 | 0.2 | 11.2 |
| ER* | 67.2 | 74.1 | 47.3 | 44.4 | 44.7 | 71 | 80.1 | 64.5 | 36.9 | 57.1 | 42.2 | 54.9 | 78.5 | 66.7 | 63.5 | 26.3 | 51.4 | 45 | 68.9 | 52.1 | 56.9 |
| ER** | 75.5 | 82 | 58.9 | 51.5 | 49.5 | 78.7 | 84.8 | 74.1 | 47.2 | 65.1 | 61 | 65.4 | 78.7 | 76.6 | 74.6 | 34.6 | 60.6 | 58.4 | 77.8 | 61.9 | 65.5 |
| ER*** | 66.7 | 68.5 | 38 | 45.7 | 40.7 | 69.8 | 81.4 | 60.3 | 30.9 | 52.3 | 67.4 | 60.4 | 81.1 | 76.6 | 79.6 | 43.9 | 61.1 | 61.3 | 77.4 | 64.7 | 61.4 |
ER* method without augmentation before uploading to memory (2500), ER** method with augmentation after uploading to memory (2500), ER*** method with augmentation after uploading to memory (1000).