five

A web Application for Bacterial Spot Classification in the Bell Pepper Plant using a Swin Transformer and Google Net Classifier a Hybrid Model

收藏
Mendeley Data2024-01-31 更新2024-06-28 收录
下载链接:
https://dataverse.harvard.edu/citation?persistentId=doi:10.7910/DVN/JPJ6SZ
下载链接
链接失效反馈
官方服务:
资源简介:
This paper addresses bacterial spot, a prevalent disease causing substantial yield losses in bell pepper plants. Effective classification of bacterial spots is crucial for farmers to implement timely precautions. The study focuses on advancing the classification of bacterial defects in bell pepper plants through a novel approach, combining the Swin Transformer and GoogleNet. The Swin Transformer, renowned for its prowess in deep learning and capturing intricate spatial relationships in images, is the foundation for feature extraction and image representation. This capability makes it adept at handling complex visual recognition tasks, such as classifying plant diseases. In conjunction, GoogleNet, known for its high classification accuracy, acts as the final layer to distinguish between healthy and infected bell pepper plants. To enhance user accessibility, a web application with an intuitive interface allows users to upload images of bell pepper leaves or plants affected by bacterial spots. Upon image submission, the Swin Transformer extracts discriminative features, which are then processed by the GoogleNet classifier to yield a binary classification decision. The introduction of real-time feedback sets this project apart, providing users with instant classification results and confidence scores. The novelty of this research lies in its groundbreaking real-time feedback mechanism, offering users immediate insights into classification outcomes and confidence levels. Achieving an impressive accuracy rate of 99 to 100 percent in testing and validation, this methodology successfully overcomes the challenge of low accuracy in pre- trained models without data augmentation. This innovative approach represents a significant advancement in plant disease classification, providing a reliable and efficient tool for farmers to combat the detrimental effects of bacterial spots on bell pepper crops.
创建时间:
2024-01-31
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作