Industrial poultry quality control dataset: multi-class classification of pathological conditions in chicken carcasses

This repository contains a dataset of 20,000 chicken carcass images collected during Marcelo dos Santos' doctoral research. The images were acquired under industrial conditions in poultry processing plants and subsequently underwent a data augmentation process. The dataset is organized into 9 folders: one containing healthy carcasses and eight representing different types of lesions. The dataset was developed to train and evaluate artificial intelligence models for automated food inspection and quality control in poultry processing. Each image captures up to three carcasses photographed on the factory production line under normal working conditions. The images are stored in JPEG format with a resolution of approximately 1.3 megapixels (1280 × 1024 pixels) and an average file size of 500 KB. Dataset structure: • 1 PDF file: dataset description and metadata • 3 JPEG files: representative image samples • Compressed folders organized by classification: - Healthy carcasses - Scratch - Arthritis - Footpad dermatitis - Dermatitis - Dermatosis - Fracture - Hematoma - Skin tear