Representation Compensation Networks for Continual Semantic Segmentation

In this work, we study the continual semantic segmentation problem, where the deep neural networks are required to incorporate new classes continually without catastrophic forgetting. We propose to use a structural re-parameterization mechanism, named representation compensation (RC) module, to deco...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	arXiv.org 2022-03
Hauptverfasser:	Chang-Bin, Zhang, Jia-Wen, Xiao, Liu, Xialei, Ying-Cong, Chen, Ming-Ming, Cheng
Format:	Artikel
Sprache:	eng
Schlagworte:	Artificial neural networks Compensation Distillation Machine learning Modules Parameterization Representations Semantic segmentation Semantics
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!