README.md

metadata

license: cc-by-4.0
task_categories:
  - object-detection
  - image-classification
  - tabular-regression
language:
  - en
tags:
  - precision-livestock-farming
  - poultry
  - robotic-inspection
  - egg-counting
  - object-detection
  - cage-aware-counting

CageAwareEggDataset

CageAwareEggDataset is a robotic inspection image dataset for cage-aware sub-cage egg counting in structured layer hen houses.

The dataset supports production-oriented egg monitoring, where egg counts need to be assigned to physical cage units rather than only reported as whole-image detections.

Dataset Description

This dataset contains robotic inspection images collected from a structured layer hen house. It includes:

• YOLO-format annotations for eggs and cage structural components
• Sub-cage-level egg-count ground truth
• Vertical-ID group split metadata for spatial generalization testing
• Ground-truth audit logs
• Final result tables for baseline and detector-variant comparisons

Annotation Classes

The object annotation classes are:

• egg
• cage_edge
• cage_roi_row1
• cage_support

Intended Use

This dataset is intended for research on:

• Precision livestock farming
• Robotic inspection in layer hen houses
• Cage-aware sub-cage egg counting
• Object detection and counting in structured agricultural environments
• Spatial generalization evaluation using vertical-ID group splits

Validation Protocols

The dataset includes two validation protocols:

1. Image-level random validation
2. Vertical-ID group split for testing spatial generalization to withheld cage positions

Data Availability

The full image archive is being uploaded. Metadata, split files, audit logs, example images, and result tables will be provided in this repository.

Citation

If you use this dataset, please cite the associated conference submission:

Cage-Aware Sub-Cage Egg Counting for Robotic Monitoring in Layer Hen Houses
Fei Shi and Weihong Ma, EPA 2026 submission.

License

This dataset is released under the CC BY 4.0 license.