Scholarlink

<p>In modern agriculture, timely diagnosis and precise classification improves crop yield and overall productivity. However, manual detection and classification of plant leaf diseases come with high computational costs and minimal performance. To address these issues, this research proposes a&#xa0;novel Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly model. This model involves data preprocessing, feature extraction, and classification phases. Initially, the Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly model carries out preprocessing utilizing the median filter, Min-Max normalization, and augmentation. After that, Deep Transfer Learning with an ensemble weight average model is applied to extract features and classify plant leaf diseases. The hyperparameters for achieving better classification performance are tuned by applying the crossover oppositional-based firefly optimization algorithm. The performances of proposed Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly is evaluated on three different plant leaf disease datasets based on diverse evaluation metrics. Also, the Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly model’s performance is compared against various state-of-the-art approaches. The Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly model surpassed the state-of-the-art approaches and achieved higher accuracy with lower computational time of 6.3 s respectively. The results confirmed that the proposed Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly model is robust for classifying various diseases from plant leaf images containing complex backgrounds.</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

Hybrid Crossover Oppositional Firefly Optimization for Enhanced Deep Transfer Learning in Plant Leaf Disease Classification

  • 期刊论文

  • 发表日期：

    2025年9月2日

查看全文
  • S. Senthil Pandi,
  • A. K. Reshmy,
  • S. Muruganandam,
  • I. Manju

摘要

In modern agriculture, timely diagnosis and precise classification improves crop yield and overall productivity. However, manual detection and classification of plant leaf diseases come with high computational costs and minimal performance. To address these issues, this research proposes a novel Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly model. This model involves data preprocessing, feature extraction, and classification phases. Initially, the Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly model carries out preprocessing utilizing the median filter, Min-Max normalization, and augmentation. After that, Deep Transfer Learning with an ensemble weight average model is applied to extract features and classify plant leaf diseases. The hyperparameters for achieving better classification performance are tuned by applying the crossover oppositional-based firefly optimization algorithm. The performances of proposed Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly is evaluated on three different plant leaf disease datasets based on diverse evaluation metrics. Also, the Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly model’s performance is compared against various state-of-the-art approaches. The Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly model surpassed the state-of-the-art approaches and achieved higher accuracy with lower computational time of 6.3 s respectively. The results confirmed that the proposed Ensemble Deep Transfer-based Hybrid Crossover Oppositional Firefly model is robust for classifying various diseases from plant leaf images containing complex backgrounds.