Scholarlink

Abstract <p>The development of an automatic system for classifying plants and their diseases at different stages of growth could play an important role in both increasing crop yields and assisting in the care of indoor plants. However, existing studies on plant and disease recognition are not systematic enough and use different datasets, making it difficult to identify the best models. In this paper, we consider the problem of constructing benchmarks for the problem of simultaneous classification of plants and their diseases and evaluate the performance of three models, MobileNetV3Small, EfficientNetB0, and DenseNet121, pretrained on the ImageNet and further trained on the PlantVillage and PlantDoc datasets. As a result of the experiments, it was found that the EfficientNetV2B0 model was the most effective for the task of plant disease recognition with an accuracy of 0.997 on the PlantVillage dataset and 0.96 on the PlantDoc dataset.</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

Recognizing Plants and Their Diseases: Benchmarks for Multiclass and Multilabel Classification

  • 期刊论文

  • 发表日期：

    2025年6月25日

查看全文
  • A. S. Shiyan,
  • I. D. Kozlov,
  • I. R. Baimuratov,
  • N. A. Zhukova

摘要

Abstract

The development of an automatic system for classifying plants and their diseases at different stages of growth could play an important role in both increasing crop yields and assisting in the care of indoor plants. However, existing studies on plant and disease recognition are not systematic enough and use different datasets, making it difficult to identify the best models. In this paper, we consider the problem of constructing benchmarks for the problem of simultaneous classification of plants and their diseases and evaluate the performance of three models, MobileNetV3Small, EfficientNetB0, and DenseNet121, pretrained on the ImageNet and further trained on the PlantVillage and PlantDoc datasets. As a result of the experiments, it was found that the EfficientNetV2B0 model was the most effective for the task of plant disease recognition with an accuracy of 0.997 on the PlantVillage dataset and 0.96 on the PlantDoc dataset.