Scholarlink

This study explores the integration of various Attention Mechanisms with pretrained ResNet and MobileNet Convolutional Neural Networks for the classification of date palm diseases. The primary objectives are to enhance the accuracy of these models, improve their interpretability by leveraging attention mechanisms to highlight relevant regions within the input data, and conduct a comprehensive comparative study. Four attention mechanisms—Squeeze Excitation, Efficient Channel Attention, Convolutional Block Attention Module, and Soft Attention—were evaluated across two datasets. The experimental results demonstrate that attention mechanisms significantly boost model performance, particularly in terms of accuracy and interpretability, while offering valuable insights from the comparative analysis across diverse models and datasets. This work has significant implications for agricultural technology, as it can lead to more precise and understandable disease detection systems, ultimately aiding farmers in making informed decisions to protect and manage their crops more effectively.

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

Enhancing Deep Learning Models for Date Palm Disease Classification Through Advanced Attention Mechanisms

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Abdelaaziz Hessane,
  • Amine El Hanafy,
  • Ahmed El Youssefi,
  • Mohamed Khalifa Boutahir,
  • El Arbi Abdellaoui Alaoui,
  • Yousef Farhaoui

摘要

This study explores the integration of various Attention Mechanisms with pretrained ResNet and MobileNet Convolutional Neural Networks for the classification of date palm diseases. The primary objectives are to enhance the accuracy of these models, improve their interpretability by leveraging attention mechanisms to highlight relevant regions within the input data, and conduct a comprehensive comparative study. Four attention mechanisms—Squeeze Excitation, Efficient Channel Attention, Convolutional Block Attention Module, and Soft Attention—were evaluated across two datasets. The experimental results demonstrate that attention mechanisms significantly boost model performance, particularly in terms of accuracy and interpretability, while offering valuable insights from the comparative analysis across diverse models and datasets. This work has significant implications for agricultural technology, as it can lead to more precise and understandable disease detection systems, ultimately aiding farmers in making informed decisions to protect and manage their crops more effectively.