Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AI-Based Teat Shape and Skin Condition Prediction for Dairy Management

Yuexing Hao, Tiancheng Yuan, Yuting Yang, Aarushi Gupta, Matthias Wieland, Ken Birman, Parminder S. Basran

TL;DR

The paper tackles automated teat health assessment in dairy farms by adapting pre-trained computer vision models to localize teats and classify teat shape and skin condition, supported by a data-processing pipeline that collects, labels, and curates keyframes from rotary milking parlors. It evaluates multiple detectors (Faster-RCNN, YOLO-F, DINO) on a small, on-farm dataset and shows that a transformer-based DINO model achieves the highest mean average precision for both teat shape (0.783) and teat skin condition (0.828), with inference times suitable for on-premise deployment on standard laptops. The work demonstrates practical, low-cost AI deployment for routine teat-health monitoring, enabling timely interventions to improve mastitis management and herd health, while acknowledging limitations like data size and label balance and outlining future expansion to additional teat-health metrics and multiple farms.

Abstract

Dairy owners spend significant effort to keep their animals healthy. There is good reason to hope that technologies such as computer vision and artificial intelligence (AI) could reduce these costs, yet obstacles arise when adapting advanced tools to farming environments. In this work, we adapt AI tools to dairy cow teat localization, teat shape, and teat skin condition classifications. We also curate a data collection and analysis methodology for a Machine Learning (ML) pipeline. The resulting teat shape prediction model achieves a mean Average Precision (mAP) of 0.783, and the teat skin condition model achieves a mean average precision of 0.828. Our work leverages existing ML vision models to facilitate the individualized identification of teat health and skin conditions, applying AI to the dairy management industry.

AI-Based Teat Shape and Skin Condition Prediction for Dairy Management

TL;DR

The paper tackles automated teat health assessment in dairy farms by adapting pre-trained computer vision models to localize teats and classify teat shape and skin condition, supported by a data-processing pipeline that collects, labels, and curates keyframes from rotary milking parlors. It evaluates multiple detectors (Faster-RCNN, YOLO-F, DINO) on a small, on-farm dataset and shows that a transformer-based DINO model achieves the highest mean average precision for both teat shape (0.783) and teat skin condition (0.828), with inference times suitable for on-premise deployment on standard laptops. The work demonstrates practical, low-cost AI deployment for routine teat-health monitoring, enabling timely interventions to improve mastitis management and herd health, while acknowledging limitations like data size and label balance and outlining future expansion to additional teat-health metrics and multiple farms.

Abstract

Dairy owners spend significant effort to keep their animals healthy. There is good reason to hope that technologies such as computer vision and artificial intelligence (AI) could reduce these costs, yet obstacles arise when adapting advanced tools to farming environments. In this work, we adapt AI tools to dairy cow teat localization, teat shape, and teat skin condition classifications. We also curate a data collection and analysis methodology for a Machine Learning (ML) pipeline. The resulting teat shape prediction model achieves a mean Average Precision (mAP) of 0.783, and the teat skin condition model achieves a mean average precision of 0.828. Our work leverages existing ML vision models to facilitate the individualized identification of teat health and skin conditions, applying AI to the dairy management industry.

Paper Structure

This paper contains 18 sections, 5 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Scientific Background
  3. Dairy Cow Teat Health Metrics
  4. Machine Learning for Dairy Health Management
  5. Data Processing
  6. Data Collection
  7. Data Labeling
  8. Automated Keyframe Selection
  9. Experimental Evaluation
  10. Model Settings
  11. Fine-tuning the Candidate Models
  12. Experimental Results
  13. Discussion
  14. Efficient Data Storage
  15. Machine Intelligence for Dairy Farms
  16. ...and 3 more sections

Figures (5)

  • Figure 1: Milking parlor and duo-camera setting illustration
  • Figure 2: Overall System Workflow
  • Figure 3: Examples of keyframes to be processed
  • Figure 4: Teat shape images, labels and train loss curve
  • Figure 5: Skin Condition bounding-box training loss curve