Related Experiment Video
Updated: May 14, 2026
09:26
Transabdominal Ultrasound for Pregnancy Diagnosis in Reeves' Muntjac Deer
Published on: January 7, 2014
10.3K
Dataset of cattle biometrics through muzzle images
Syed Umaid Ahmed1, Jaroslav Frnda2, Muhammad Waqas1
1National University of Computer and Emerging Sciences FAST-NUCES, Karachi, Pakistan.
Data in Brief
|February 19, 2024
Summary
A new Cattle Biometrics Dataset with over 8,000 annotated images aids research in cattle recognition and management. This resource supports advancements in computer vision and machine learning for agriculture.
Area of Science:
- Agriculture
- Computer Science
- Data Science
- Biometrics
Background:
- Biometric recognition in cattle is crucial for management and identification.
- Existing datasets may lack diversity in breeds, ages, and environmental conditions.
- Muzzle prints are a reliable biometric characteristic for cattle.
Purpose of the Study:
- To introduce a comprehensive and diverse Cattle Biometrics Dataset.
- To facilitate research in cattle biometric recognition and computer vision applications.
- To support machine learning model development, particularly transfer learning.
Main Methods:
- Collected over 8,000 annotated cattle photographs from public markets and farms.
- Ensured diversity in breeds, ages, genders, and environmental conditions.
- Annotated images with a focus on the muzzle region as a key biometric feature.
Main Results:
- Established a large-scale, annotated dataset of cattle biometrics.
- Dataset includes images from various camera qualities and conditions.
- Provides a valuable resource for interdisciplinary research and development.
Conclusions:
- The Cattle Biometrics Dataset is a significant resource for advancing cattle recognition and management.
- It enables algorithmic development in computer vision and machine learning.
- Promotes interdisciplinary collaboration for innovation in agriculture and data science.
