ISIC Dataset

该数据集是由国际皮肤成像协作组织（ISIC）提供的开放获取的皮肤镜图像分析基准挑战，旨在对黑色素瘤进行分类。它包含了来自ISIC数据库的带注释的皮肤病变照片。该数据集混合了恶性和良性皮肤病变的图像，所有图像都被调整为224×224×3像素的大小。具体规模上，训练集包含6996张图像，测试集包含3000张图像。任务是对黑色素瘤进行分类。

This dataset is an open-access dermoscopic image analysis benchmark challenge provided by the International Skin Imaging Collaboration (ISIC), which aims to classify melanoma. It contains annotated skin lesion photos from the ISIC database. This dataset includes images of both malignant and benign skin lesions, and all images have been resized to 224×224×3 pixels. In terms of scale, the training set consists of 6996 images, while the test set contains 3000 images. The task of this challenge is to classify melanoma.

搜集汇总

背景与挑战

背景概述

ISIC数据集专注于皮肤病变检测，特别是黑色素瘤的识别。该数据集支持知识蒸馏方法的研究，旨在开发轻量级但高精度的模型，适用于临床环境。