# 🖼️ Instance Segmentation with Mask R-CNN (ResNet-50 + FPN) using Detectron2
Today, I successfully ran an instance segmentation model using Mask R-CNN with the ResNet-50 backbone and Feature Pyramid Network (FPN), based on the config file:
mask_rcnn_R_50_FPN_3x.yaml.
🔍 Model Architecture Overview
- ResNet-50: Backbone network to extract rich feature representations from the image.
- FPN (Feature Pyramid Network): Improves feature maps at multiple scales for better detection of small and large objects.
- Mask R-CNN: Builds on top of Faster R-CNN by adding a segmentation branch to predict masks at the pixel level.
✅ Key Learnings & Workflow
- Understood how to use and modify model config files in Detectron2.
- Explored the model loading process from pretrained checkpoints.
- Ran inference successfully on a sample input and verified the output.
Top comments (1)
good