Welcome to the Broiler Datasets Repository - a collection of curated datasets focused on broiler detection, broiler behavior, pose estimation, and re-identification. These datasets were used to develop and evaluate deep learning models for precision livestock farming applications, including video classification and pose estimation. Our published research leverages these datasets to advance automated broiler behavior analysis.
The datasets were created to support research in:
- Automated detection and classification of broiler behaviors.
- Development of video recognition models for precise behavior quantification.
- Pose estimation and movement analysis of broilers for health monitoring.
- Broiler re-identification across multiple views for individual tracking.
- Improving deep learning techniques in precision livestock farming applications.
| Dataset Name | Description | README File |
|---|---|---|
| Lameness Pose Dataset | Images and keypoint annotations for broiler lameness pose estimation | Pose Estimation README |
| Behavior Recognition | Videos categorized into normal, preening, and stretching behaviors | Behavior Recognition README |
We have an ongoing effort to expand these datasets with more annotated data to improve the robustness and accuracy of broiler behavior and identification models.
- For each dataset, please refer to its detailed README file in .md format, which includes download instructions and usage details.
- You can download and explore the datasets individually.
- Feel free to raise issues or contribute!
- Name: Hao Gan -- Amin Nasiri
- Email: [email protected] -- [email protected]
- Affiliation: Department of Biosystems Engineering and Soil Science, The University of Tennessee, Knoxville, TN, USA
MIT License.
Contributions are welcome!