Human Activity Recognition on Wallhack LOS Biquad 1.8k (test)
94.87Accuracy (Test Set)Ensemble (All 5)
Evaluation results
| Method | Links | |
|---|---|---|
| Ensemble (All 5)Method=Ensemble, Components=Models 1, 2, 3, 4, 5, Image Size=224 x 224 x 32026.06 | 94.87 | |
| Model 3 (MobileNetV2)Architecture=MobileNetV2, Batch Size=32, Mode=fine-tuned, Image Size=224 x 224 x 32026.06 | 94.21 | |
| Model 1 (Deep CNN)Architecture=Custom Deep CNN, Batch Size=16, Image Size=224 x 224 x 32026.06 | 92.34 | |
| Model 5 (EfficientNetB0)Architecture=EfficientNetB0, Batch Size=32, Mode=fine-tuned, Image Size=224 x 224 x 32026.06 | 91.56 | |
| Model 2 (Wide CNN)Architecture=Custom Wide CNN, Batch Size=16, Image Size=224 x 224 x 32026.06 | 88.78 | |
| Model 4 (ResNet50V2)Architecture=ResNet50V2, Batch Size=32, Mode=fine-tuned, Image Size=224 x 224 x 32026.06 | 87.65 |