Two-in-One Refinement for Interactive Segmentation
Abstract
Deep convolutional neural networks are now mainstream for click-based interactive image segmentation. Most frameworks refine false negatives and false positive regions via a succession of positive and negative clicks placed centrally in these regions. We propose a simple yet intuitive two-in-one refinement strategy placing clicks on the boundary of the object of interest. As boundary clicks are a strong cue for extracting the object of interest and we find that they are much more effective in correcting wrong segmentation masks. In addition, we propose a boundary-aware loss that encourages segmentation masks to respect instance boundaries. We place our new refinement scheme and loss formulation within a task-specialized segmentation framework and achieve state-of-the-art performance on the standard datasets - Berkeley, Pascal VOC 2012, DAVIS, and MS COCO. We exceed competing methods by 6.5%, 9.4%, 10.5% and 2.5%, respectively.
Bibtex
@INPROCEEDINGS{majumder-2020-two, author = {Majumder, Soumajit and Rai, Abhinav and Khurana, Ansh and Yao, Angela}, title = {Two-in-One Refinement for Interactive Segmentation}, booktitle = {Proc. 31st British Machine Vision Conference (BMVC20)}, year = {2020}, abstract = {Deep convolutional neural networks are now mainstream for click-based interactive image segmentation. Most frameworks refine false negatives and false positive regions via a succession of positive and negative clicks placed centrally in these regions. We propose a simple yet intuitive two-in-one refinement strategy placing clicks on the boundary of the object of interest. As boundary clicks are a strong cue for extracting the object of interest and we find that they are much more effective in correcting wrong segmentation masks. In addition, we propose a boundary-aware loss that encourages segmentation masks to respect instance boundaries. We place our new refinement scheme and loss formulation within a task-specialized segmentation framework and achieve state-of-the-art performance on the standard datasets - Berkeley, Pascal VOC 2012, DAVIS, and MS COCO. We exceed competing methods by 6.5%, 9.4%, 10.5% and 2.5%, respectively.} }