October 2, 2025Open Access

HCTNet: Hybrid CNN--Mamba Network for Real-Time Semantic Segmentation in Urban Traffic Scenes

View Full Paper

Ask AI

Bookmark

HCTNet achieves 81.0% mean Intersection-over-Union (mIoU) at 60.5 FPS in urban traffic scenes, showcasing high efficiency.
Integration of a Convolutional State Module (CSM) enhances the receptive field, contributing to improved segmentation outcomes.
A shared decoder streamlines predictions from multiple branches, maintaining real-time performance while ensuring model coherence.
On ApolloScape, HCTNet attains a 73.8% mIoU, demonstrating its applicability across different urban scene datasets.

Discussion

No takes yet. Share an insight, caveat, or question.

Proposes HCTNet to enhance real-time semantic segmentation, achieving 81% mIoU in urban settings, indicating improved context capture.

Meng et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark