Instance Segmenation_balloons
Computer Vision, Instance Segmentation, 2024
This project involves training a custom dataset using Facebook Research’s Detectron2 framework, specifically with the Mask R-CNN model.
The dataset used for training is the balloons dataset, which was also employed by Facebook Research when introducing Detectron2. The project demonstrates how to adapt and train Detectron2 for custom object detection tasks. The code and images are available in this GitHub repository.
Link to the gihub repository: