GalliformeSpectra: A Hen Breed Dataset

Published: 25 October 2023| Version 1 | DOI: 10.17632/nk3zbvd5h8.1
Contributors:
,

Description

This dataset includes a variety of low-resolution images showcasing ten globally recognized hen breeds. These breeds were carefully chosen from different parts of the world to ensure a diverse and comprehensive representation. The dataset serves as a visual resource, offering a detailed depiction of the unique characteristics of these hen breeds, which aids in their accurate classification. It consists of ten distinct categories: Bielefeld, Blackorpington, Brahma, Buckeye, Fayoumi, Leghorn, Newhampshire, Plymouthrock, Sussex, and Turken, comprising a total of 1010 original JPG images that were later resized and converted to PNG format. After applying augmentation techniques, the total number of images increased to 5050. The dataset is organized into three variations: one with original images, another with resized images, and a third with augmented images. Each variation is further divided into ten separate folders, each dedicated to a specific hen breed. The images vary in size and have been subjected to data augmentation, which is essential for training machine vision deep learning models. Augmentation includes transformations like left and right flips, random adjustments in brightness (0.7-1.1), and random zoom (90% area). Consequently, an additional 505 augmented images were created from the original images in each category, resulting in a dataset comprising a total of 5050 augmented images (505 per category).

Files

Institutions

Bangladesh University of Business and Technology, University of Greenwich, Universiti Sains Malaysia

Categories

Image Processing, Machine Learning, Computer, Agricultural Animal, Deep Learning

Licence