Table 2 Segmentation details for YOLOv8 training, including dataset size, training split percentage, and number of epochs for each segmented facial part across the three datasets.
From: A segment-based framework for explainability in animal affective computing
Dataset | Segment | Dataset size for YOLOv8 | Dataset split - training | Input img size | #epochs |
|---|---|---|---|---|---|
Cat pain | Face | 54 | 0.85 | 800x800 | 10 |
Cat pain | Ears | 54 | 0.85 | 224x224 | 25 |
Cat pain | Eyes | 54 | 0.85 | 224x224 | 25 |
Cat pain | Mouth | 54 | 0.85 | 224x224 | 25 |
Horse pain | Face | 39 | 0.82 | 224x224 | 20 |
Horse pain | Ears | 39 | 0.82 | 224x224 | 35 |
Horse pain | Eyes | 39 | 0.82 | 224x224 | 35 |
Horse pain | Mouth | 39 | 0.82 | 224x224 | 20 |
Dog emotion | Face | 135 | 0.87 | 640x640 | 30 |
Dog emotions | Ears | 145 | 0.87 | 224x224 | 10 |
Dog emotions | Eyes | 145 | 0.87 | 224x224 | 10 |
Dog emotions | Mouth | 145 | 0.87 | 224x224 | 20 |
