Table 4 Shelf detection results after training models from scratch and fine-tuning pretrained models on the proposed dataset.
From: Real-time retail planogram compliance application using computer vision and virtual shelves
Model | Precision | Recall | F1 Score | mAP@50 |
|---|---|---|---|---|
YOLOv8 | 98.88% | 98.94% | 98.87% | 99.32% |
Deformable DETR | 97.12% | 98.03% | 97.73% | 98.72% |
Pretr. YoloV8 | 99.23% | 98.93% | 99.18% | 99.41% |
Pretr. Deformable DETR | 81.02% | 87.27% | 84.03% | 98.71% |
