Cattle Body Parts Image Dataset for Object Detection

Intro

This dataset is a curated collection of images featuring various cattle body parts aimed at facilitating object detection tasks. The dataset contains a total of 428 high-quality photos, meticulously annotated with three distinct classes: "Back," "Head," and "Leg."

The dataset can be downloaded using this link. The dataset is also available at Roboflow Universe.

A YOLOv7X model has been trained using the dataset, and you can access the trained weights through this link.

Motivation

Accurate and reliable identification of different cattle body parts is crucial for various agricultural and veterinary applications. This dataset aims to provide a valuable resource for researchers, developers, and enthusiasts working on object detection tasks involving cattle, ultimately contributing to advancements in livestock management, health monitoring, and related fields.

Data

Overview

• Total Images: 428
• Classes: Back, Head, Leg
• Annotations: Bounding boxes for each class

Below is an example image from the dataset.

Contents

📦 Cattle_Body_Parts_OD.zip
 ┣ 📂 images
 ┃  ┣ 📜 image1.jpg
 ┃  ┣ 📜 image2.jpg
 ┃  ┗ ...
 ┗ 📂 annotations
    ┣ 📜 image1.json
    ┣ 📜 image2.json
    ┗ ...

Annotation Format

Each annotation file corresponds to an image in the dataset and is formatted as per the LabelMe JSON standard. These annotations define the bounding box coordinates for each labeled body part, enabling straightforward integration into object detection pipelines.

License

This work is licensed under a Creative Commons Attribution 4.0 International License.

Disclaimer

This dataset has been collected from publicly available sources. I do not claim ownership of the data and have no intention of infringing on any copyright. The material contained in this dataset is copyrighted to their respective owners. I have made every effort to ensure the data is accurate and complete, but I cannot guarantee its accuracy or completeness. If you believe any data in this dataset infringes on your copyright, please get in touch with me immediately so I can take appropriate action.

Contact

For any questions, concerns, or collaborations, please don't hesitate to contact me on my LinkedIn.