AI Paper Atlas

Research map / Detection and segmentation / Weakly supervised detection and segmentation

Tendency-driven Mutual Exclusivity for Weakly Supervised Incremental Semantic Segmentation

Chongjie Si, Xuehui Wang, Xiaokang Yang, Wei Shen

ECCV 2024 · Detection and segmentation › Weakly supervised detection and segmentation · approach: Weakly supervised semantic segmentation · 4 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add tendency-driven mutual exclusivity (TME) to Teddy (ablation)Teddy · 15-5 VOC (overlap), all classesmIoU: 68.9 -> 72.0Improved
add TME to a common WILSS model (no SAM binarization, no fusion) (ablation)WILSS baseline · 15-5 VOC (overlap), all classesmIoU: 67.0 -> 70.1Improved
add SAM-based prediction binarization for old classes in TME (ablation)Teddy · 15-5 VOC (overlap), all classesmIoU: 70.1 -> 70.7Improved
add SAM-based prediction fusion for new classes (ablation)Teddy · 10-2 VOC (overlap), all classesmIoU: 40.8 -> 43.1Improved
optimize fusion coefficients U and V instead of fixed constants (ablation)Teddy · 10-2 VOC (overlap), all classesmIoU: 40.9 -> 43.1Improved
Teddy versus WILSON, new classesTeddy · 15-5 VOC (overlap), classes 16-20mIoU: 41.7 -> 51.4Improved
Teddy versus WILSON in multi-step settingTeddy · 10-2 VOC (overlap), all classesmIoU: 32.5 -> 43.1Improved
Teddy versus WILSON on 10-10 VOCTeddy · 10-10 VOC (overlap), all classesmIoU: 65.0 -> 66.5Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy