Mask R-CNN
At a glance
Mask R-CNN identifies objects and traces each one's outline in an image.
In one paragraph
Mask R-CNN identifies objects and outlines the pixels belonging to each one, distinguishing separate objects even when they share a category. It extends Faster R-CNN with an additional outline-prediction step and also demonstrates how the approach can locate human body keypoints.
Authors 4
- Kaiming He
- Georgia Gkioxari
- Piotr Dollár
- Ross Girshick
