CNFOOD-241

该数据集名为CNFOOD-241，是公开可用的食品数据集中图片尺寸最为统一的一个，包含了近二十万张尺寸均为600×600的图片。该数据集被划分为训练集（119,514张图片）、验证集（51,354张图片）和测试集（20,943张图片）。数据集的归一化熵显示，类别之间存在较大不平衡性，这增加了其挑战性。总的来说，该数据集涵盖了大约200,000张图片，分布在241个类别中，其任务是进行图像分类。

The dataset designated as CNFOOD-241 is one of the publicly available food datasets with the most consistent image dimensions. It contains nearly 200,000 images, all with a fixed resolution of 600×600. The dataset is split into three subsets: a training set with 119,514 images, a validation set with 51,354 images, and a test set with 20,943 images. The normalized entropy of the dataset reveals significant class imbalance, which increases the difficulty of the task. Overall, this dataset covers approximately 200,000 images across 241 categories, and its target task is image classification.

搜集汇总

背景与挑战

背景概述

CNFOOD-241是一个用于细粒度食品分类的数据集，与Res-VMamba模型结合使用，支持深度残差学习和选择性状态空间模型的研究。数据集图像列表可在指定文件夹中找到，原始数据来源于Mendeley平台。