Related Experiment Video
Updated: Jul 5, 2026
06:08
Quantitative Visualization and Detection of Skin Cancer Using Dynamic Thermal Imaging
Published on: May 5, 2011
17.2K
Deep learning enhanced thermographic modeling for early and precise mastitis detection in Sahiwal cows
S L Gayathri1, M Bhakat1, T K Mohanty1
1Livestock Production Management Division, ICAR-National Dairy Research Institute, Karnal 132001, Haryana, India.
Research in Veterinary Science
|September 19, 2025
Summary
Thermal imaging and deep learning accurately detect mastitis in dairy cows. This precision dairy farming approach aids in early identification of sub-clinical mastitis (SCM) and clinical mastitis (CM), improving livestock health management.
Area of Science:
- Veterinary Medicine
- Animal Science
- Artificial Intelligence in Agriculture
Background:
- Mastitis is a significant dairy production disease requiring advanced diagnostic tools.
- Current methods for mastitis detection can be labor-intensive and may not always provide early warnings.
- Integrated and precision-based diagnostic approaches are crucial for effective dairy herd management.
Purpose of the Study:
- To evaluate the efficacy of thermal imaging combined with deep learning for mastitis detection in Sahiwal cows.
- To develop and assess Convolutional Neural Network (CNN) models for classifying udder quarters as healthy, Sub-clinical Mastitis (SCM), or Clinical Mastitis (CM).
- To enhance precision dairy farming through improved mastitis diagnostic capabilities.
Main Methods:
- Capturing thermal images of the udder region in lactating Sahiwal cows using a handheld thermal camera.
- Classifying udder quarters based on California Mastitis Test (CMT) scores, Somatic Cell Count (SCC) values, and thermal image analysis.
- Developing and training CNN models to differentiate between healthy and mastitis-affected udder quarters (CM and SCM).
Main Results:
- A CNN model distinguishing healthy quarters from Clinical Mastitis (CM) achieved 99% accuracy, precision, recall, and F1-score.
- A CNN model differentiating healthy quarters from Sub-clinical Mastitis (SCM) demonstrated 84% testing accuracy, 0.87 precision, 0.79 recall, and 0.83 F1-score.
- Thermal imaging coupled with CNN shows high potential for accurate and efficient mastitis detection.
Conclusions:
- CNN-based thermal imaging offers a promising, non-invasive method for accurate mastitis detection in dairy cows.
- This technology can significantly contribute to advancements in precision dairy farming and proactive livestock health management.
- Early and precise diagnosis of mastitis, including SCM and CM, can lead to timely interventions and improved animal welfare.
