Scholarlink

<p>Infrared–visible fusion alleviates the limitations of single-modality detection, especially under low illumination and haze. However, most existing methods rely on sequential or parallel attention fusion, which limits representation ability and efficiency. To address these issues, we propose DASYOLO, a cross-modal object detector that combines shallow feature enhancement with dual-attention-guided cross-modal fusion. Specifically, DASYOLO incorporates a BiAttention module to strengthen shallow feature representations through complementary channel and spatial attention, and a Dual Attention Synergy (DAS) module that jointly models channel importance and spatial saliency through multiplicative coupling. This design improves cross-modal feature discrimination while maintaining high inference efficiency. Extensive evaluations on M3FD, LLVIP, and our UGVLQ dataset demonstrate that DASYOLO achieves 83.9%&#xa0;mAP 0.5/% on M3FD (+ 3.2%) and 95.2%&#xa0;mAP 0.5/% on LLVIP (+ 0.8%). With only 7.52&#xa0;M parameters and an inference speed of 143&#xa0;FPS, DASYOLO significantly outperforms parameter-intensive alternatives such as ICAF, achieving a favorable balance between detection accuracy and efficiency.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

DASYOLO: dual-attention-synergistic YOLO for cross-modality object detection

  • 期刊论文

  • 发表日期：

    2026年4月25日

查看全文
  • Yunjia Yang,
  • Xiaoxia Wang,
  • Fengbao Yang,
  • Weiwei Du,
  • Bo Li

摘要

Infrared–visible fusion alleviates the limitations of single-modality detection, especially under low illumination and haze. However, most existing methods rely on sequential or parallel attention fusion, which limits representation ability and efficiency. To address these issues, we propose DASYOLO, a cross-modal object detector that combines shallow feature enhancement with dual-attention-guided cross-modal fusion. Specifically, DASYOLO incorporates a BiAttention module to strengthen shallow feature representations through complementary channel and spatial attention, and a Dual Attention Synergy (DAS) module that jointly models channel importance and spatial saliency through multiplicative coupling. This design improves cross-modal feature discrimination while maintaining high inference efficiency. Extensive evaluations on M3FD, LLVIP, and our UGVLQ dataset demonstrate that DASYOLO achieves 83.9% mAP 0.5/% on M3FD (+ 3.2%) and 95.2% mAP 0.5/% on LLVIP (+ 0.8%). With only 7.52 M parameters and an inference speed of 143 FPS, DASYOLO significantly outperforms parameter-intensive alternatives such as ICAF, achieving a favorable balance between detection accuracy and efficiency.