Scholarlink

The poultry industry is a critical global source of food and economic stability. Despite its significance, the poultry industry faces persistent threats from diseases, which can result in significant economic losses and potential public health risks. Accurate and fast identification of infections can help poultry farmers choose appropriate treatment, saving economic loss and ensuring public health. In this regard, a deep learning model can help to identify the poultry diseases early. In this work, a transfer learning model ResNet-9 is proposed for the detection of poultry diseases like Salmonella, New Castle Disease (NCD) and Coccidiosis. The suggested model achieves 98.83% accuracy on the Kaggle Poultry Diseases Detection dataset. A mobile application is developed that enables the poulterer to easily identify the diseases by uploading the image of the chicken’s faecal. The suggested model’s performance is compared to other current research, demonstrating that it outperforms previous results.

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

A Smartphone-Based Deep Learning Framework for Detection and Classification of Poultry Diseases from Faecal Images

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • P. Kaviya,
  • S. Sibi Siddharthan,
  • M. Kishore,
  • M. Muthuram

摘要

The poultry industry is a critical global source of food and economic stability. Despite its significance, the poultry industry faces persistent threats from diseases, which can result in significant economic losses and potential public health risks. Accurate and fast identification of infections can help poultry farmers choose appropriate treatment, saving economic loss and ensuring public health. In this regard, a deep learning model can help to identify the poultry diseases early. In this work, a transfer learning model ResNet-9 is proposed for the detection of poultry diseases like Salmonella, New Castle Disease (NCD) and Coccidiosis. The suggested model achieves 98.83% accuracy on the Kaggle Poultry Diseases Detection dataset. A mobile application is developed that enables the poulterer to easily identify the diseases by uploading the image of the chicken’s faecal. The suggested model’s performance is compared to other current research, demonstrating that it outperforms previous results.