Scholarlink

<p>Floods pose significant threats to both human life and infrastructure, necessitating timely and accurate monitoring for effective disaster management. Advancements in computer vision and machine learning have enabled the development of sophisticated systems for flood classification using imagery. In this study, we introduce MoSWIN, a novel framework that integrates Monkey Search Optimization (MSO) with the Swin Transformer to enhance the accuracy and robustness of flood classification models. The Swin Transformer, with its hierarchical architecture and shifted window mechanism, excels at capturing both local and global features in flood-related imagery. MSO is employed to extract high-level features from input images, facilitating improved feature engineering. Subsequently, the Swin Transformer performs classification tasks on the features extracted through MSO. Our proposed MoSWIN model outperforms several existing state-of-the-art approaches, achieving an accuracy of 96.53%. This indicates a significant improvement in image classification and optimization performance, leading to more effective flood detection. Experimental results demonstrate that MoSWIN can accurately distinguish between flooded and non-flooded areas, surpassing the performance of conventional methods.</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

MoSWIN: Monkey Search Optimization and SWIN-based Flood Classification Architecture

  • 期刊论文

  • 发表日期：

    2025年6月13日

查看全文
  • Vinay Dubey,
  • Rahul Katarya

摘要

Floods pose significant threats to both human life and infrastructure, necessitating timely and accurate monitoring for effective disaster management. Advancements in computer vision and machine learning have enabled the development of sophisticated systems for flood classification using imagery. In this study, we introduce MoSWIN, a novel framework that integrates Monkey Search Optimization (MSO) with the Swin Transformer to enhance the accuracy and robustness of flood classification models. The Swin Transformer, with its hierarchical architecture and shifted window mechanism, excels at capturing both local and global features in flood-related imagery. MSO is employed to extract high-level features from input images, facilitating improved feature engineering. Subsequently, the Swin Transformer performs classification tasks on the features extracted through MSO. Our proposed MoSWIN model outperforms several existing state-of-the-art approaches, achieving an accuracy of 96.53%. This indicates a significant improvement in image classification and optimization performance, leading to more effective flood detection. Experimental results demonstrate that MoSWIN can accurately distinguish between flooded and non-flooded areas, surpassing the performance of conventional methods.