Work

Variation-aware proxy learning for semantic segmentation

Year: 2025

Type: article

Abstract: In semantic segmentation, accurately modeling intra-class variance is essential for capturing fine-grained details and resolving ambiguity near class boundaries. While existing proxy-based embedding methods represent each class with a single prototype, they struggle to reflect diverse intra-class structures, especially in complex scenes. In this paper, we propose a novel representation learning framework called Variation-Aware Proxy Learning, which introduces a representative proxy to encode shared class semantics and multiple variation vectors to capture fine-grained intra-class variations. These components are integrated through a factorized similarity score, enabling more expressive and discriminative embedding structures. To further enhance learning in ambiguous regions, we introduce focal modulation and design a new Compositional Similarity Loss composed of attraction and repulsion terms that adaptively amplify the contribution of hard examples. Our method is model-agnostic and requires no additional inference-time cost. Extensive experiments across multiple segmentation benchmarks—Cityscapes, COCO-Stuff10k, and iSAID—and diverse backbones including CNNs and Transformers, demonstrate consistent improvements in mIoU and boundary precision, particularly in challenging regions with high intra-class variability.

Source: Neurocomputing

DOI: 10.1016/j.neucom.2025.131783

Authors: Haejun Bae, Byung Cheol Song

Institution: Inha University

Language: English

Topic: Advanced Neural Network Applications

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Estimated APC paid: $2,670