Semi-supervised semantic segmentation meets masked modeling : Fine-grained locality learning matters in consistency regularization

Wentao Pan

doi:10.1016/j.patcog.2025.112586

Semi-supervised semantic segmentation meets masked modeling : Fine-grained locality learning matters in consistency regularization

2025-10-15

0

PRE

AI

DOI：10.1016/j.patcog.2025.112586

Original

Original request for help

Save

Abstract

En 中文

• Fine-grained locality learning for semi-supervised semantic segmentation. • The task-specific mask modeling facilitates fine-grained locality perception. • The multi-scale ensembling strategy improves pseudo-label quality. • Superior performance on benchmarks, and plug-and-play flexibility.

Semi-supervised semantic segmentation meets masked modeling : Fine-grained locality learning matters in consistency regularization

Abstract

Journal

Organization