ERSGNet: An Edge-Guided Residual Spatial Gating Fusion Network for Real-Time Semantic Segmentation
Open Access
Abstract
With the development of autonomous driving and general real-time vision applications, semantic segmentation is increasingly expected to provide reliable pixel-level scene understanding under strict latency and computational constraints. Existing real-time semantic segmentation methods still struggle to effectively balance contextual semantics and local details during feature fusion. This difficulty may lead to boundary blur and the loss of fine-grained structures, and is mainly related to insufficient selectivity and boundary friendliness in the fusion process. To address this issue, the Edge-Guided Residual Spatial Gating Fusion Network (ERSGNet)is proposed as a three-branch decoupled framework for real-time semantic segmentation. In the ERSGNet, detail features , contextual features , and edge features are separately extracted to reduce mutual interference among different types of information. Spatially Gated Interaction (SGI) is introduced to inject contextual semantics into the detail branch in a pixel-wise selective manner, the Parallel Context Pyramid (PCP) is used to enhance multi-scale contextual representation, and Edge-Guided Residual Spatial Gating Fusion (ERSG-Fuse) performs edge-guided adaptive fusion between and under the guidance of . The ERSGNet is evaluated on the CamVid and Cityscapes datasets. On CamVid, the ERSGNet achieves 73.33% mean Intersection-over-Union (mIoU) at 166.27 frames per second (FPS). On Cityscapes, the ERSGNet achieves 69.21% mIoU at 53.93 FPS. These results suggest that the ERSGNet achieves a favorable accuracy–efficiency trade-off while supporting boundary-related and fine-grained structural representation.
Keywords
Funding Information
- National Natural Science Foundation of China (62403112)
- National Natural Science Foundation of China (52527810)
- China Postdoctoral Science Foundation (2025T180470)
- Fundamental Research Funds for the Central Universities (N2404025)
- Natural Science Foundation of Liaoning Province (2024-BSBA-19)
This publication has 25 references indexed in Scilit:
- DBBGNet: a dual-branch network with boundary information guidance for real-time semantic segmentationThe Visual Computer, 2026
- Taguchi-optimized DeepLabV3+ for semantic segmentation in autonomous driving applicationsAin Shams Engineering Journal, 2026
- AEAFFNet: enhancing real-time semantic segmentation through attention-enhanced adaptive feature fusionThe Journal of Supercomputing, 2026
- HARD: Hardware-Aware Lightweight Real-Time Semantic Segmentation Model Deployable from Edge to GPU2024
- Real-time semantic segmentation for autonomous driving: A review of CNNs, Transformers, and BeyondJournal of King Saud University Computer and Information Sciences, 2024
- EBUNet: a fast and accurate semantic segmentation network with lightweight efficient bottleneck unitComplex & Intelligent Systems, 2023
- LAANet: lightweight attention-guided asymmetric network for real-time semantic segmentationNeural Computing and Applications, 2022
- BiSeNet V2: Bilateral Network with Guided Aggregation for Real-Time Semantic SegmentationInternational Journal of Computer Vision, 2021
- BiSeNet: Bilateral Segmentation Network for Real-Time Semantic Segmentation2018
- Semantic object classes in video: A high-definition ground truth databasePattern Recognition Letters, 2009