SCMF-Net: Sparse Self-Attention Driven Cross-Modal Fusion for Robust Detection in Complex Road Scenes
IEEE Sensors Journal · 01.04.2026 · 26 (7) · 10721-10730
DOI: 10.1109/jsen.2026.3664865 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Sensors Journal, erschienen am 01.04.2026. Verfasst von Yunze He und 6 weiteren Autor:innen. Die Arbeit wurde bisher 14-mal zitiert (Crossref) und nennt 33 Literaturangaben. Gefördert von: Science and Technology Innovation Leader of Hunan Province, Basic and Applied Basic Research Foundation of Guangdong Province.
Abstract (Originalsprache)
This article introduces sparse cross-modal fusion network (SCMF-Net), a lightweight multimodal perception framework designed to enhance representation quality and inference efficiency while minimizing computational overhead. To address the sparsity and irregular distribution of LiDAR point clouds, an intensity-aware depth encoding strategy is proposed to enhance the structural cues in the depth modality. Additionally, a dual-branch backbone is employed to further strengthen feature extraction. Building upon this, feature fusion local self-attention (FFLSA) is introduced to enable efficient cross-modal fusion. FFLSA leverages self-attention clustering (SAC) to identify salient cross-modal regions, and self-attention fusion and purification (SAFP) to refine feature aggregation and reduce redundancy, forming an effective region-selection-refined fusion mechanism. Additionally, a cross-modal feature fusion module (FFM) is proposed, which jointly models spatial and channel attention to enable adaptive RGB–depth interaction with fine-grained weighting. Extensive experiments on the KITTI dataset and a custom RGB–LiDAR benchmark under challenging conditions (fog, low light, and overexposed) validate the effectiveness of the proposed approach. SCMF-Net achieves 79.3% mean average precision (mAP) on our dataset and 88.3% on KITTI, surpassing current state-of-the-art detection methods.
Förderung
- Science and Technology Innovation Leader of Hunan Province
- Basic and Applied Basic Research Foundation of Guangdong Province
Zitieren
APA
He, Y., Hao, Y., Qian, M., Deng, B., Zhang, L., Cheng, L., & Wang, Y. (2026). SCMF-Net: Sparse Self-Attention Driven Cross-Modal Fusion for Robust Detection in Complex Road Scenes. IEEE Sensors Journal, 26(7), 10721-10730. https://doi.org/10.1109/jsen.2026.3664865
Vancouver
He Y, Hao Y, Qian M, Deng B, Zhang L, Cheng L, et al. SCMF-Net: Sparse Self-Attention Driven Cross-Modal Fusion for Robust Detection in Complex Road Scenes. IEEE Sensors Journal. 2026;26(7):10721-10730. doi:10.1109/jsen.2026.3664865
BibTeX
@article{He2026,
author = {He, Yunze and Hao, Yousheng and Qian, Mengying and Deng, Baoyuan and Zhang, Lilian and Cheng, Liang and Wang, Yaonan},
title = {SCMF-Net: Sparse Self-Attention Driven Cross-Modal Fusion for Robust Detection in Complex Road Scenes},
journal = {IEEE Sensors Journal},
year = {2026},
volume = {26},
number = {7},
pages = {10721--10730},
doi = {10.1109/jsen.2026.3664865}
}
EUVIMED