This is my project repo for ECE800
- This project is a review mainly in the object detection filed. Traditional approaches include region selection, feature extraction, and classifier. It introduces object detection based on R-CNN and illustrates why and how the R-CNN family evolved. It also covers Mask R-CNN, which has a great performance on high-quality objects segmentation task.
- In this repo, there are two ipynb files. Rcnn.ipynb is a simple R-CNN implementation. Its dataset is also included in this repo. Detectron2.ipynb is a conclusion of how to use the R-CNN models in detectron2. But you need to use COCO dataset or prepare the custom dataset by yourself.
The final report is here. You can also access the ECE800 course website for more information.