A computer algorithm for scoring cow teats based on udder images

In this work, we aimed to develop a computer algorithm for detecting and scoring cow teats according to udder images. In addition, by assessing three teat traits, i.e. the length, thickness, and placements, we scored them according to international standards. By applying the algorithm to 71 cows, we...

Full description

Saved in:
Bibliographic Details
Published in:New Zealand journal of agricultural research Vol. 68; no. 7; pp. 1696 - 1706
Main Authors: Chuah, Chong Sheng, Nordbø, Øyvind, Ho, Harvey
Format: Journal Article
Language:English
Published: Taylor & Francis 02.12.2025
Subjects:
ISSN:0028-8233, 1175-8775, 1175-8775
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this work, we aimed to develop a computer algorithm for detecting and scoring cow teats according to udder images. In addition, by assessing three teat traits, i.e. the length, thickness, and placements, we scored them according to international standards. By applying the algorithm to 71 cows, we were able to detect all four teats in 62 of 71 (85%) cow udder images, and differentiate the front and rear teats in 56 of 62 (92%) udder images. Based on the three teat traits, we scored the teats of 56 cows. The final teat score was given by averaging the three scores for each cow. We discuss the limitations of the algorithm, and also the issues with teat conformation and image-taking itself, e.g. teats not facing perpendicular to the camera. In conclusion, we have developed the first computerised teat scoring system based on cow udder images. Further improvements require tuning parameters of the workflow and the image-taking process.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0028-8233
1175-8775
1175-8775
DOI:10.1080/00288233.2024.2412684