Maturity-Aware Active Learning for Semantic Segmentation with Hierarchically-Adaptive Sample Assessment

TL;DR

MADBAL introduces a hierarchical, multiview active learning approach for semantic segmentation that adaptively assesses sample importance, significantly improving early-stage performance and reducing annotation effort.

Contribution

It proposes MADBAL, a novel hierarchical active learning method with a maturity-aware uncertainty measure for semantic segmentation, outperforming existing methods.

Findings

Outperforms state-of-the-art on Cityscapes and PASCAL VOC datasets.

Achieves significant performance gains in early active learning stages.

Reduces training and annotation effort effectively.

Abstract

Active Learning (AL) for semantic segmentation is challenging due to heavy class imbalance and different ways of defining "sample" (pixels, areas, etc.), leaving the interpretation of the data distribution ambiguous. We propose "Maturity-Aware Distribution Breakdown-based Active Learning'' (MADBAL), an AL method that benefits from a hierarchical approach to define a multiview data distribution, which takes into account the different "sample" definitions jointly, hence able to select the most impactful segmentation pixels with comprehensive understanding. MADBAL also features a novel uncertainty formulation, where AL supporting modules are included to sense the features' maturity whose weighted influence continuously contributes to the uncertainty detection. In this way, MADBAL makes significant performance leaps even in the early AL stage, hence reducing the training burden significantly. It outperforms state-of-the-art methods on Cityscapes and PASCAL VOC datasets as verified in our extensive experiments.