Saved in:
Bibliographic Details
Main Authors: Caglayan, Ali, Imamoglu, Nevrez, Kouyama, Toru
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.15705
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908781091225600
author Caglayan, Ali
Imamoglu, Nevrez
Kouyama, Toru
author_facet Caglayan, Ali
Imamoglu, Nevrez
Kouyama, Toru
contents This work focuses on national-scale land-use/land-cover (LULC) semantic segmentation using ALOS-2 single-polarization (HH) SAR data over Japan, together with a companion binary water detection task. Building on SAR-W-MixMAE self-supervised pretraining [1], we address common SAR dense-prediction failure modes, boundary over-smoothing, missed thin/slender structures, and rare-class degradation under long-tailed labels, without increasing pipeline complexity. We introduce three lightweight refinements: (i) injecting high-resolution features into multi-scale decoding, (ii) a progressive refine-up head that alternates convolutional refinement and stepwise upsampling, and (iii) an $α$-scale factor that tempers class reweighting within a focal+dice objective. The resulting model yields consistent improvements on the Japan-wide ALOS-2 LULC benchmark, particularly for under-represented classes, and improves water detection across standard evaluation metrics.
format Preprint
id arxiv_https___arxiv_org_abs_2601_15705
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Enhanced LULC Segmentation via Lightweight Model Refinements on ALOS-2 SAR Data
Caglayan, Ali
Imamoglu, Nevrez
Kouyama, Toru
Computer Vision and Pattern Recognition
This work focuses on national-scale land-use/land-cover (LULC) semantic segmentation using ALOS-2 single-polarization (HH) SAR data over Japan, together with a companion binary water detection task. Building on SAR-W-MixMAE self-supervised pretraining [1], we address common SAR dense-prediction failure modes, boundary over-smoothing, missed thin/slender structures, and rare-class degradation under long-tailed labels, without increasing pipeline complexity. We introduce three lightweight refinements: (i) injecting high-resolution features into multi-scale decoding, (ii) a progressive refine-up head that alternates convolutional refinement and stepwise upsampling, and (iii) an $α$-scale factor that tempers class reweighting within a focal+dice objective. The resulting model yields consistent improvements on the Japan-wide ALOS-2 LULC benchmark, particularly for under-represented classes, and improves water detection across standard evaluation metrics.
title Enhanced LULC Segmentation via Lightweight Model Refinements on ALOS-2 SAR Data
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2601.15705