RSC：525种鸟类图像分类数据集

该数据集是专门用于鸟类图像分类研究的大规模标准化数据集，包含525种不同鸟类物种的高质量彩色图像，总计89,885张图像。数据集包括84,635张训练图像、2,625张测试图像和2,625张验证图像。所有图像均为224×224×3像素的JPG格式，经过严格的数据清洗和质量保证流程处理。每张图像中只包含一只鸟类，且鸟类通常占据至少50%的像素区域，图像质量优良且未经人工增强。数据集通过互联网搜索收集原始图像，随后进行去重处理、质量筛选和标准化处理，移除了重复、近似重复以及有缺陷的低信息图像。该数据集适用于深度学习模型训练和鸟类自动识别研究。

This is a large-scale standardized dataset specifically developed for bird image classification research. It contains high-quality color images of 525 distinct bird species, with a total of 89,885 images. The dataset is split into 84,635 training images, 2,625 test images, and 2,625 validation images. All images are in JPG format with a resolution of 224×224×3 pixels, and have undergone strict data cleaning and quality assurance procedures. Each image contains exactly one bird, and the bird usually occupies at least 50% of the pixel area. The images are of excellent quality and have not been artificially enhanced. The original images were collected through internet searches, followed by deduplication, quality screening and standardization processing, with duplicate, near-duplicate and defective low-information images removed. This dataset is applicable to deep learning model training and bird automatic recognition research.

搜集汇总

背景与挑战

背景概述

该数据集是包含525种鸟类物种的大规模标准化图像分类数据集，总计89,885张高质量彩色图像，图像尺寸统一为224×224×3像素，适用于深度学习模型训练和鸟类自动识别研究。