Scholarlink

<p>Currently, the mainstream restoration tasks under adverse weather conditions have predominantly focused on single-weather scenarios. However, in reality, multiple weather conditions always coexist and their degree of mixing is usually unknown. Under such complex and diverse weather conditions, single-weather restoration models struggle to meet practical demands. This is particularly important in areas such as autonomous driving, where there is an urgent need for models that can effectively handle mixed weather conditions and enhance image quality in an automated manner. In this paper, we propose a multi-weather image restoration method which includes a task-sequence generator module that combined with the task intra-patch block to effectively extract task-specific features embedded in degraded images. The task intra-patch block introduces an external learnable sequence that promotes the network in capturing task-specific information. Additionally, we introduce a histogram-based transformer module as the backbone of our network, enabling the capture of global and local dynamic-range features. Our proposed model achieves state-of-the-art performance on public datasets. The code will be available at <a href="https://github.com/iaslay/Removeweather.git">https://github.com/iaslay/Removeweather.git</a>.</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

Multi-weather Image Restoration via Histogram-Based Transformer Feature Enhancement

  • 期刊论文

  • 发表日期：

    2025年8月20日

查看全文
  • Yang Wen,
  • Anyu Lai,
  • Bo Qian,
  • Wuzhen Shi,
  • Wenming Cao

摘要

Currently, the mainstream restoration tasks under adverse weather conditions have predominantly focused on single-weather scenarios. However, in reality, multiple weather conditions always coexist and their degree of mixing is usually unknown. Under such complex and diverse weather conditions, single-weather restoration models struggle to meet practical demands. This is particularly important in areas such as autonomous driving, where there is an urgent need for models that can effectively handle mixed weather conditions and enhance image quality in an automated manner. In this paper, we propose a multi-weather image restoration method which includes a task-sequence generator module that combined with the task intra-patch block to effectively extract task-specific features embedded in degraded images. The task intra-patch block introduces an external learnable sequence that promotes the network in capturing task-specific information. Additionally, we introduce a histogram-based transformer module as the backbone of our network, enabling the capture of global and local dynamic-range features. Our proposed model achieves state-of-the-art performance on public datasets. The code will be available at https://github.com/iaslay/Removeweather.git.