Segment unknown objects using transformers by stopping the gradient using SWORD
Segment unknown objects using transformers by stopping the gradient using SWORD
Exploring Transformers for Open-world Instance Segmentation
arXiv paper abstract https://arxiv.org/abs/2308.04206
arXiv PDF paper https://arxiv.org/pdf/2308.04206.pdf
Open-world instance segmentation is a rising task, which aims to segment all objects in the image by learning from a limited number of base-category objects.
… utilize the Transformer for open-world instance segmentation and present SWORD … introduce to attach the stop-gradient operation before classification head and further add IoU heads for discovering novel objects.
… demonstrate that a simple stop-gradient operation not only prevents the novel objects from being suppressed as background, but also allows the network to enjoy the merit of heuristic label assignment.
… propose a novel contrastive learning framework to enlarge the representations between objects and background.
… maintain a universal object queue to obtain the object center, and dynamically select positive and negative samples from the object queries for contrastive learning.
… models achieve state-of-the-art performance in various open-world cross-category and cross-dataset generalizations …
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
