EGRNet: A Lightweight Semantic Segmentation Network with Edge-Gated Refinement and Adversarial Sensing

2026-07-21

AI Summary

This work addresses the demand for low-cost, high-accuracy, and robust real-time semantic segmentation in autonomous driving and smart city applications by proposing EGRNet, a lightweight neural network. The architecture integrates depthwise separable convolutions and dilated residual blocks to reduce computational complexity while capturing multi-scale contextual information. It introduces an Edge-Gated Refinement (EGR) module that adaptively fuses features to enhance boundary details, alongside a lightweight adversarial-aware mechanism to improve robustness. With only 0.46 million parameters, EGRNet achieves a mean Intersection-over-Union (mIoU) of 65.28% on the Cityscapes benchmark, striking an excellent balance among accuracy, efficiency, and adversarial robustness, thereby demonstrating strong suitability for edge deployment.

Abstract

As autonomous systems and smart cities continue to evolve, the demand for efficient and robust scene understanding becomes increasingly critical. Semantic segmentation plays a key role in enabling autonomous vehicles to comprehend complex urban environments. However, achieving high accuracy with minimal computational cost remains a significant challenge. In this paper, we present Edge-Gated Refinement Network (EGRNet), a lightweight and efficient deep learning model designed for real-time semantic segmentation in urban scenarios. The model incorporates depthwise separable convolutions to reduce computational complexity and dilated residual blocks for capturing rich multi-scale contextual information. Additionally, we introduce a novel Edge-Gated Refinement (EGR) module, which adaptively fuses original and refined features through a learnable gating mechanism, enhancing boundary preservation and edge-sensitive regions. To further improve feature representation, Squeeze-and-Excitation (SE) attention is applied across the network. With only 0.46M parameters, EGRNet achieves state-of-the-art performance while maintaining low computational overhead. When evaluated on the Cityscapes dataset, the model attains a mean Intersection over Union (mIoU) of 65.28%, demonstrating strong accuracy with minimal resource consumption. Moreover, we introduce a lightweight adversarial attack detection strategy, ensuring robustness against adversarial inputs without compromising real-time performance. By combining efficiency, accuracy, and resilience, EGRNet is well-suited for deployment on edge devices in safety-critical real-time applications.

Problem

• semantic segmentation
• lightweight network
• real-time performance
• edge devices
• adversarial robustness

Innovation

• Edge-Gated Refinement
• lightweight semantic segmentation
• adversarial sensing
• depthwise separable convolution
• squeeze-and-excitation attention