Table 7 Classification performance of different network models on the satiety dataset.
From: Fish feeding behavior recognition via lightweight two stage network and satiety experiments
Model | Params (M) | Accuracy (%) |
|---|---|---|
C2D | 24.30 (+ 7.4%) | 76.8 (− 21.7%) |
GoogleNet | 59.8 (+ 165.8%) | 99.1 (+1%) |
ResNet34 | 212.9 (+ 846.2%) | 96.2 (− 1.9%) |
EfficientNetv2 | 202.6 (+ 800.4%) | 97.4 (− 0.7%) |
MobileVitv3 | 415.3 (+ 1745.8%) | 98.6 (+ 0.5%) |
ShuffleNetv2 | 12.6 (− 44.0%) | 53.7 (− 45.3%) |
Densenet | 8.7 (− 61.3%) | 91.0 (− 7.2%) |
Ours (baseline) | 22.5 | 98.1 |
