Table 3 Detection results of Faster R-CNN based on different models
From: Effective automated pipeline for 3D reconstruction of synapses based on deep learning
Dataset | EM | Size | Model | AP | Rate |
|---|---|---|---|---|---|
(A) Cortex | ATUM-SEM | 1000×1000 | ZF | 82.0% | 9 fps |
| Â | Â | Â | VGG16 | 83.2% | 3 fps |
| Â | Â | Â | ResNet50 | 84.1% | 4 fps |
(B) Hippocampus | FIB-SEM | 500×500 | ZF | 86.8% | 36 fps |
| Â | Â | Â | VGG16 | 87.4% | 12 fps |
| Â | Â | Â | ResNet50 | 90.9% | 16 fps |