Authors:
R. Dineshraj, S. Sindhu, Bushra Rehman

Addresses:
Department of Data Science and Business Systems, School of Computing, College of Engineering and Technology, SRM Institute of Science and Technology, Kattankulathur, Chennai, Tamil Nadu, India. Institute of Pathology and Diagnostic Medicine, Khyber Medical University, Peshawar, Khyber Pakhtunkhwa, Pakistan.

Abstract:

The paper is vital to detecting Mastitis in dairy cows at an early stage to enhance the health of dairy animals and dairy production. This paper presents an exploratory data analysis (EDA) of a dataset on Cow and Mastitis Detection acquired from Roboflow to assess its applicability to computer vision-based detection models. The data comprises annotated images of cow activities, such as feeding, drinking, lying, and standing, and health status, such as healthy udder, lumpy infected cow, and mastitis-infected udder. The analysis consists of an overview of the dataset, validation of the training, validation, and testing splits, and visualization of sample images with YOLO-format annotations to analyze the quality of the data. The analysis of class distribution indicates a considerable imbalance among the seven classes, which can affect model performance. Analysis of image dimensions shows differences in resolution, and the distribution of objects per image sheds light on annotation density. Also, the annotations are divided into health and behavior groups to assess the datasets' focus. The quality checks of the annotation also detect inconsistencies, such as missing or sparse labels. Altogether, the EDA provides essential information about the structure and challenges of the data, creating a solid foundation for powerful AI-based systems for mastitis detection.

Keywords: Mastitis Detection; Dairy Cows; Exploratory Data Analysis (EDA); Computer Vision; Object Detection; Class Imbalance; Animal Health Monitoring; Dairy Production.

Received on: 05/09/2024, Revised on: 26/11/2024, Accepted on: 09/04/2025, Published on: 07/05/2026

DOI: 10.69888/FTSASS.2026.000648

FMDB Transactions on Sustainable Applied Sciences, 2026 Vol. 3 No. 1, Pages: 35-46

• Views : 36
• Downloads : 10

Download PDF