A$^2$LC: Active and Automated Label Correction for Semantic Segmentation

2025-06-13

AI Summary

To address the high cost and error-prone nature of pixel-level manual annotation in semantic segmentation, this paper proposes a two-stage label correction framework integrating *active learning* and *automatic propagation*. The method jointly leverages foundation-model-derived pseudo-labels, uncertainty estimation, and human feedback to design an adaptive query function that balances class distribution and correction efficacy. Crucially, it introduces the first large-scale, propagation-based automatic correction mechanism for *non-queried* samples—overcoming a fundamental limitation of conventional active label correction (ALC), which restricts correction exclusively to queried instances. Evaluated on Cityscapes, the framework achieves state-of-the-art performance under a 20% labeling budget, improving mIoU by 27.23%. On PASCAL VOC 2012, it significantly enhances both correction efficiency and generalization capability.

Technology Category

• Machine Learning: Active Learning
• Computer Vision: Segmentation
• Search and Optimization: Learning to Search

Application Category

• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Active Label Correction (ALC) has emerged as a promising solution to the high cost and error-prone nature of manual pixel-wise annotation in semantic segmentation, by selectively identifying and correcting mislabeled data. Although recent work has improved correction efficiency by generating pseudo-labels using foundation models, substantial inefficiencies still remain. In this paper, we propose Active and Automated Label Correction for semantic segmentation (A$^2$LC), a novel and efficient ALC framework that integrates an automated correction stage into the conventional pipeline. Specifically, the automated correction stage leverages annotator feedback to perform label correction beyond the queried samples, thereby maximizing cost efficiency. In addition, we further introduce an adaptively balanced acquisition function that emphasizes underrepresented tail classes and complements the automated correction mechanism. Extensive experiments on Cityscapes and PASCAL VOC 2012 demonstrate that A$^2$LC significantly outperforms previous state-of-the-art methods. Notably, A$^2$LC achieves high efficiency by outperforming previous methods using only 20% of their budget, and demonstrates strong effectiveness by yielding a 27.23% performance improvement under an equivalent budget constraint on the Cityscapes dataset. The code will be released upon acceptance.

Problem

• Reducing manual annotation costs in semantic segmentation
• Improving label correction efficiency using automated methods
• Enhancing performance for underrepresented tail classes

Innovation

• Automated label correction using annotator feedback
• Adaptively balanced acquisition for tail classes
• Integration of automated stage in ALC pipeline