早期准备有回报：用于类增量语义分割的新分类器预调优
Early Preparation Pays Off: New Classifier Pre-tuning for Class Incremental Semantic Segmentation
AI总结：
本文提出NeST新分类器预调优方法，通过学习旧分类器到新分类器的变换来初始化新分类器，防止特征提取器剧烈变化，并在Pascal VOC 2012和ADE20K上显著提升类增量语义分割性能。
摘要：
类增量语义分割旨在学习新任务的同时保留旧知识，然而，它受到灾难性遗忘和背景偏移问题的阻碍。先前的研究表明，初始化新分类器至关重要，并主要集中于从背景分类器迁移知识或为未来类别准备分类器，而忽视了新分类器的灵活性和方差。在本文中，我们提出了一种在正式训练过程之前应用的新分类器预调优（NeST）方法，该方法学习从旧分类器到生成新分类器的变换以用于初始化，而不是直接调优新分类器的参数。我们的方法可以使新分类器与骨干网络对齐并适应新数据，防止在学习新类别时特征提取器发生剧烈变化。此外，我们设计了一种考虑跨任务类别相似性的策略来初始化变换中使用的矩阵，有助于实现稳定性-可塑性权衡。在Pascal VOC 2012和ADE20K数据集上的实验表明，所提出的策略可以显著提高先前方法的性能。代码可在https://github.com/zhengyuan-xie/ECCV24_NeST获取。
英文摘要：
Class incremental semantic segmentation aims to preserve old knowledge while learning new tasks, however, it is impeded by catastrophic forgetting and background shift issues. Prior works indicate the pivotal importance of initializing new classifiers and mainly focus on transferring knowledge from the background classifier or preparing classifiers for future classes, neglecting the flexibility and variance of new classifiers. In this paper, we propose a new classifier pre-tuning~(NeST) method applied before the formal training process, learning a transformation from old classifiers to generate new classifiers for initialization rather than directly tuning the parameters of new classifiers. Our method can make new classifiers align with the backbone and adapt to the new data, preventing drastic changes in the feature extractor when learning new classes. Besides, we design a strategy considering the cross-task class similarity to initialize matrices used in the transformation, helping achieve the stability-plasticity trade-off. Experiments on Pascal VOC 2012 and ADE20K datasets show that the proposed strategy can significantly improve the performance of previous methods. The code is available at \url{https://github.com/zhengyuan-xie/ECCV24_NeST}.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
