Mitigating Background Shift in Class-Incremental Semantic Segmentation
Class-Incremental Semantic Segmentation(CISS) aims to learn new classes without forgetting the old ones, using only the labels of the new classes. To achieve this, two popular strategies are employed: 1) pseudo-labeling and knowledge distillation to preserve prior knowledge; and 2) background weight transfer, which leverages the broad coverage of background in learning new classes by transferring background weight to the new class classifier. However, the first strategy heavily relies on the old model in detecting old classes while undetected pixels are regarded as the background, thereby leading to the background shift towards the old classes(i.e., misclassification of old class as background). Additionally, in the case of the second approach, initializing the new class classifier with background knowledge triggers a similar background shift issue, but towards the new classes. To address these issues, we propose a background-class separation framework for CISS. To begin with, selective pseudo-labeling and adaptive feature distillation are to distill only trustworthy past knowledge. On the other hand, we encourage the separation between the background and new classes with a novel orthogonal objective along with label-guided output distillation. Our state-of-the-art results validate the effectiveness of these proposed methods.
Code (1)
Tasks
Class Incremental LearningClass-Incremental Semantic SegmentationContinual LearningContinual Semantic SegmentationDisjoint 15-1Disjoint 15-5Disjoint 19-1Knowledge DistillationOverlapped 100-10Overlapped 100-5Overlapped 100-50Overlapped 10-1Overlapped 15-1Overlapped 15-5Overlapped 19-1Overlapped 50-50Overlapped 5-3SegmentationSemantic SegmentationMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Attribution-aware Weight Transfer: A Warm-Start Initialization for Class-Incremental Semantic Segmentation
In class-incremental semantic segmentation (CISS), deep learning architectures suffer from the critical problems of catastrophic forgetting and semantic background shift. Although recent works focused on these issues, ex…
Class-Incremental Semantic SegmentationContinual LearningOverlapped 100-10Overlapped 100-5+8Background Adaptation with Residual Modeling for Exemplar-Free Class-Incremental Semantic Segmentation
Class Incremental Semantic Segmentation~(CISS), within Incremental Learning for semantic segmentation, targets segmenting new categories while reducing the catastrophic forgetting on the old categories.Besides, backgroun…
Class-Incremental Semantic SegmentationExemplar-FreeIncremental LearningKnowledge Distillation+2Causes of Catastrophic Forgetting in Class-Incremental Semantic Segmentation
Class-incremental learning for semantic segmentation (CiSS) is presently a highly researched field which aims at updating a semantic segmentation model by sequentially learning new semantic classes. A major challenge in …
class-incremental learningClass Incremental LearningClass-Incremental Semantic SegmentationIncremental Learning+3Modeling the Background for Incremental Learning in Semantic Segmentation
Despite their effectiveness in a wide range of tasks, deep architectures suffer from some important limitations. In particular, they are vulnerable to catastrophic forgetting, i.e. they perform poorly when they are requi…
Continual LearningDisjoint 10-1Disjoint 15-1Disjoint 15-5+8Gradient-Semantic Compensation for Incremental Semantic Segmentation
Incremental semantic segmentation aims to continually learn the segmentation of new coming classes without accessing the training data of previously learned classes. However, most current methods fail to address catastro…
SegmentationSemantic Segmentation