Robust Trust Region for Weakly Supervised Segmentation

TL;DR

This paper introduces a robust trust region method for weakly supervised segmentation that leverages strong low-level solvers for regularized losses, significantly enhancing training quality in scenarios with limited pixel labels.

Contribution

The paper presents a novel trust region approach enabling the use of advanced regularizers with strong optimization methods in deep learning for weakly supervised segmentation.

Findings

Achieved state-of-the-art results on weakly supervised segmentation tasks.

Demonstrated the effectiveness of integrating strong low-level solvers into neural network training.

Improved segmentation accuracy with limited pixel annotations.

Abstract

Acquisition of training data for the standard semantic segmentation is expensive if requiring that each pixel is labeled. Yet, current methods significantly deteriorate in weakly supervised settings, e.g. where a fraction of pixels is labeled or when only image-level tags are available. It has been shown that regularized losses - originally developed for unsupervised low-level segmentation and representing geometric priors on pixel labels - can considerably improve the quality of weakly supervised training. However, many common priors require optimization stronger than gradient descent. Thus, such regularizers have limited applicability in deep learning. We propose a new robust trust region approach for regularized losses improving the state-of-the-art results. Our approach can be seen as a higher-order generalization of the classic chain rule. It allows neural network optimization to use strong low-level solvers for the corresponding regularizers, including discrete ones.