Evidential Deep Learning for Class-Incremental Semantic Segmentation

TL;DR

This paper introduces a novel evidential deep learning approach for class-incremental semantic segmentation, effectively modeling unlabeled classes and managing background-shift issues to improve performance on incremental learning benchmarks.

Contribution

It proposes a new method using Dirichlet distributions to model class evidence and background uncertainty, addressing unlabeled classes and background-shift in incremental segmentation.

Findings

Outperforms state-of-the-art on Pascal VOC and ADE20k benchmarks.

Effectively models unlabeled classes and background uncertainty.

Maintains performance over multiple incremental learning steps.

Abstract

Class-Incremental Learning is a challenging problem in machine learning that aims to extend previously trained neural networks with new classes. This is especially useful if the system is able to classify new objects despite the original training data being unavailable. While the semantic segmentation problem has received less attention than classification, it poses distinct problems and challenges since previous and future target classes can be unlabeled in the images of a single increment. In this case, the background, past and future classes are correlated and there exist a background-shift. In this paper, we address the problem of how to model unlabeled classes while avoiding spurious feature clustering of future uncorrelated classes. We propose to use Evidential Deep Learning to model the evidence of the classes as a Dirichlet distribution. Our method factorizes the problem into a separate foreground class probability, calculated by the expected value of the Dirichlet distribution, and an unknown class (background) probability corresponding to the uncertainty of the estimate. In our novel formulation, the background probability is implicitly modeled, avoiding the feature space clustering that comes from forcing the model to output a high background score for pixels that are not labeled as objects. Experiments on the incremental Pascal VOC, and ADE20k benchmarks show that our method is superior to state-of-the-art, especially when repeatedly learning new classes with increasing number of increments.