VegFru

VegFru是用于细粒度视觉分类的特定领域数据集。VegFru根据蔬菜和水果的食用特征对其进行分类，每个图像都包含至少一个烹饪用途相同的蔬菜或水果的可食用部分。特别是，所有图像都被分层标记。当前版本涵盖了25个上层类别和292个下属类别的蔬菜和水果。它总共包含160,000多个图像，每个从属类至少包含200个图像。

VegFru is a domain-specific dataset for fine-grained visual classification. It categorizes vegetables and fruits based on their edible traits, where each image features at least one edible portion of a vegetable or fruit with identical culinary use. Notably, all images are hierarchically labeled. The current version covers 25 upper-level categories and 292 subordinate-level categories of vegetables and fruits, comprising over 160,000 images in total, with each subordinate category containing no fewer than 200 images.

搜集汇总

背景与挑战

背景概述

VegFru是一个细粒度视觉分类数据集，专注于蔬菜和水果的分类，包含25个上层类别和292个下属类别，总计超过160,000张图像。数据集由中国科学技术大学于2017年发布，适用于图像检索和色彩恒常性等计算机视觉任务。