Animal Activity Recognition on horse dataset
93.42AccuracyCS_CE
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| CS_CEmethod_full_name=Cost-sensitive Cross-Entropy loss2026.04 | 93.42 | 79.54 | 81.72 | 80.03 | |
| Baselineloss function=standard cross-entropy loss, sampling=single sampling rate2026.04 | 93.32 | 79.5 | 81.78 | 79.74 | |
| ACSLmethod_full_name=Adaptive Class Suppression Loss2026.04 | 92.49 | 77.28 | 78.81 | 76.4 | |
| IBA-Net2026.04 | 92.22 | 84.34 | 84.05 | 84.65 | |
| CB_FLmethod_full_name=Class-balanced Focal Loss2026.04 | 91.4 | 83.64 | 83.27 | 84.21 | |
| KMSmethod_full_name=KMeansSMOTE2026.04 | 90.15 | 83.07 | 82.97 | 84.27 | |
| RUSmethod_full_name=RandomUnderSampler2026.04 | 85.37 | 80.25 | 79.28 | 85.16 |