The objective is to use Mask RCNN technique for instance segmentation on the COCO dataset.
The COCO dataset is a large image dataset designed for Object Detection, Segmentation, person keypoints detection, and Caption Generation. COCO has several features:
- Object segmentation
- Recognition in context
- Superpixel stuff segmentationo330K images (>200K labeled)
- 1.5 million object instances
- 80 object categories
- 91 stuff categories
- 5 captions per image
- 250,000 people with keypoints
•Clone the Facebook Detectron Modelusing the following link:
•Load the pre-trained weights using the following link:
•Use the Mask RCNN Architecture and the pre-trained weights to generate predictions for our own images or images from the COCO dataset
•Visualize the Results
- The input I provided:
- The output image with masks