Slightly Shift New Classes to Remember Old Classes for Video Class-Incremental Learning

arXiv:2404.00901v1 Announce Type: new
Abstract: Recent video class-incremental learning usually excessively pursues the accuracy of the newly seen classes and relies on memory sets to mitigate catastrophic forgetting of the old classes. However, limited storage only allows storing a few representative videos. So we propose SNRO, which slightly shifts the features of new classes to remember old classes. Specifically, SNRO contains Examples Sparse(ES) and Early Break(EB). ES decimates at a lower sample rate to build memory sets and uses interpolation …

abstract accuracy arxiv catastrophic forgetting class cs.cv however incremental memory shift storage type video videos