Poultry Pathology Visual Dataset

Empowering Poultry Care Through Visual Analysis

About Dataset

The chicken ecology image File is a comprehensive collection of chicken fecal images curated and augmented to aid small to medium-scale chicken farmers in the early detection and diagnosis of chicken diseases.

Key Features

Data Source: Images obtained using the Open Data Kit (ODK) app on mobile phones, sourced from various repositories.

Image Dimensions: Resized to a uniform size of 224×224 pixels.

Classes: Categorized into four distinct classes: Coccidiosis, Healthy, Newcastle Disease, Salmonella.

Data Augmentation: Enhanced using techniques like cropping, affine transformations, blurring, noise addition, and more.

Data Split: Divided into “train” (100,000 images per class), “val” (10,000 images per class), and “test” sets.

Performance Results

Achieved a peak accuracy of 97.62% on the test set and 97.84% on the validation set using Vision Transformer models.

Papers Published

“Poultry Disease Identification in Fecal Images Using Vision Transformer” by Jayavrinda Vrindavanam et al. (Medicon Agriculture & Environmental Sciences, 2024).

Usage

Ideal for researchers, veterinarians, and poultry farmers developing automated systems for disease detection and management.

Can be used for disease classification, severity assessment, and disease progression tracking.

Availability

The dataset is publicly available and can be accessed through specified directories, with proper attribution appreciated when used for research or application development.