CherryChèvre: A fine-grained dataset for goat detection in natural environments

Area of Science:

• Computer Vision
• Machine Learning
• Agricultural Science

Background:

• Accurate goat detection is crucial for precision agriculture, animal welfare, and husbandry.
• Existing datasets may lack diversity in environmental conditions, limiting algorithm generalization.

Purpose of the Study:

• Introduce a novel, large-scale dataset for goat detection.
• Provide a benchmark for evaluating machine learning algorithms in agricultural contexts.

Main Methods:

• Collected 6160 images of goats under diverse environmental conditions.
• Expert annotators ensured high accuracy and consistency in image labeling.
• Dataset is made publicly available for research and development.

Main Results:

• The dataset comprises 6160 meticulously annotated images.
• Images capture goats in various settings, crucial for robust model training.
• The dataset serves as a standardized benchmark for goat detection algorithms.

Conclusions:

• This dataset significantly advances computer vision research in agriculture.
• Facilitates the development of improved goat monitoring and management systems.
• Enables further research into animal behavior analysis and welfare.