MFFP-Net: Building Segmentation in Remote Sensing Images via Multi-Scale Feature Fusion and Foreground Perception Enhancement
The accurate segmentation of small target buildings in high-resolution remote sensing images remains challenging due to two critical issues: (1) small target buildings often occupy few pixels in complex backgrounds, leading to frequent background confusion, and (2) significant intra-class variance c...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-05-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/17/11/1875 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1850161038049148928 |
|---|---|
| author | Huajie Xu Qiukai Huang Haikun Liao Ganxiao Nong Wei Wei |
| author_facet | Huajie Xu Qiukai Huang Haikun Liao Ganxiao Nong Wei Wei |
| author_sort | Huajie Xu |
| collection | DOAJ |
| description | The accurate segmentation of small target buildings in high-resolution remote sensing images remains challenging due to two critical issues: (1) small target buildings often occupy few pixels in complex backgrounds, leading to frequent background confusion, and (2) significant intra-class variance complicates feature representation compared to conventional semantic segmentation tasks. To address these challenges, we propose a novel Multi-Scale Feature Fusion and Foreground Perception Enhancement Network (MFFP-Net). This framework introduces three key innovations: (1) a Multi-Scale Feature Fusion (MFF) module that hierarchically aggregates shallow features through cross-level connections to enhance fine-grained detail preservation, (2) a Foreground Perception Enhancement (FPE) module that establishes pixel-wise affinity relationships within foreground regions to mitigate intra-class variance effects, and (3) a Dual-Path Attention (DPA) mechanism combining parallel global and local attention pathways to jointly capture structural details and long-range contextual dependencies. Experimental results demonstrate that the IoU of the proposed method achieves improvements of 0.44%, 0.98% and 0.61% compared to mainstream state-of-the-art methods on the WHU Building, Massachusetts Building, and Inria Aerial Image Labeling datasets, respectively, validating its effectiveness in handling small targets and intra-class variance while maintaining robustness in complex scenarios. |
| format | Article |
| id | doaj-art-f1a75a981bb447bf8c331ed82645f006 |
| institution | OA Journals |
| issn | 2072-4292 |
| language | English |
| publishDate | 2025-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj-art-f1a75a981bb447bf8c331ed82645f0062025-08-20T02:22:59ZengMDPI AGRemote Sensing2072-42922025-05-011711187510.3390/rs17111875MFFP-Net: Building Segmentation in Remote Sensing Images via Multi-Scale Feature Fusion and Foreground Perception EnhancementHuajie Xu0Qiukai Huang1Haikun Liao2Ganxiao Nong3Wei Wei4Guangxi Key Laboratory of Multimedia Communications and Network Technology, School of Computer and Electronic Information, Guangxi University, Nanning 530004, ChinaGuangxi Key Laboratory of Multimedia Communications and Network Technology, School of Computer and Electronic Information, Guangxi University, Nanning 530004, ChinaGuangxi Key Laboratory of Multimedia Communications and Network Technology, School of Computer and Electronic Information, Guangxi University, Nanning 530004, ChinaGuangxi Key Laboratory of Multimedia Communications and Network Technology, School of Computer and Electronic Information, Guangxi University, Nanning 530004, ChinaSchool of Computer Science and Technology, Huazhong University of Science and Technology, Wuhan 430070, ChinaThe accurate segmentation of small target buildings in high-resolution remote sensing images remains challenging due to two critical issues: (1) small target buildings often occupy few pixels in complex backgrounds, leading to frequent background confusion, and (2) significant intra-class variance complicates feature representation compared to conventional semantic segmentation tasks. To address these challenges, we propose a novel Multi-Scale Feature Fusion and Foreground Perception Enhancement Network (MFFP-Net). This framework introduces three key innovations: (1) a Multi-Scale Feature Fusion (MFF) module that hierarchically aggregates shallow features through cross-level connections to enhance fine-grained detail preservation, (2) a Foreground Perception Enhancement (FPE) module that establishes pixel-wise affinity relationships within foreground regions to mitigate intra-class variance effects, and (3) a Dual-Path Attention (DPA) mechanism combining parallel global and local attention pathways to jointly capture structural details and long-range contextual dependencies. Experimental results demonstrate that the IoU of the proposed method achieves improvements of 0.44%, 0.98% and 0.61% compared to mainstream state-of-the-art methods on the WHU Building, Massachusetts Building, and Inria Aerial Image Labeling datasets, respectively, validating its effectiveness in handling small targets and intra-class variance while maintaining robustness in complex scenarios.https://www.mdpi.com/2072-4292/17/11/1875remote sensingbuilding segmentationfeature fusionforeground perceptiondual-path attention |
| spellingShingle | Huajie Xu Qiukai Huang Haikun Liao Ganxiao Nong Wei Wei MFFP-Net: Building Segmentation in Remote Sensing Images via Multi-Scale Feature Fusion and Foreground Perception Enhancement Remote Sensing remote sensing building segmentation feature fusion foreground perception dual-path attention |
| title | MFFP-Net: Building Segmentation in Remote Sensing Images via Multi-Scale Feature Fusion and Foreground Perception Enhancement |
| title_full | MFFP-Net: Building Segmentation in Remote Sensing Images via Multi-Scale Feature Fusion and Foreground Perception Enhancement |
| title_fullStr | MFFP-Net: Building Segmentation in Remote Sensing Images via Multi-Scale Feature Fusion and Foreground Perception Enhancement |
| title_full_unstemmed | MFFP-Net: Building Segmentation in Remote Sensing Images via Multi-Scale Feature Fusion and Foreground Perception Enhancement |
| title_short | MFFP-Net: Building Segmentation in Remote Sensing Images via Multi-Scale Feature Fusion and Foreground Perception Enhancement |
| title_sort | mffp net building segmentation in remote sensing images via multi scale feature fusion and foreground perception enhancement |
| topic | remote sensing building segmentation feature fusion foreground perception dual-path attention |
| url | https://www.mdpi.com/2072-4292/17/11/1875 |
| work_keys_str_mv | AT huajiexu mffpnetbuildingsegmentationinremotesensingimagesviamultiscalefeaturefusionandforegroundperceptionenhancement AT qiukaihuang mffpnetbuildingsegmentationinremotesensingimagesviamultiscalefeaturefusionandforegroundperceptionenhancement AT haikunliao mffpnetbuildingsegmentationinremotesensingimagesviamultiscalefeaturefusionandforegroundperceptionenhancement AT ganxiaonong mffpnetbuildingsegmentationinremotesensingimagesviamultiscalefeaturefusionandforegroundperceptionenhancement AT weiwei mffpnetbuildingsegmentationinremotesensingimagesviamultiscalefeaturefusionandforegroundperceptionenhancement |