2020 SIIM-ISIC Melanoma Classification challenge dataset

本数据集名为‘2020 SIIM-ISIC Melanoma Classification challenge dataset’，由纪念斯隆-凯特琳癌症中心等全球多个中心合作创建。数据集包含33,126张皮肤镜图像，来自2056名患者，平均每位患者有16个病变图像。数据集的创建过程涉及从多个临床图像数据库中查询和筛选，确保图像质量和诊断标签的准确性。该数据集主要用于支持机器学习挑战，特别是在评估和诊断黑色素瘤方面，旨在通过提供患者级别的上下文信息，帮助改善黑色素瘤的诊断准确性和效率。

This dataset, named '2020 SIIM-ISIC Melanoma Classification Challenge Dataset', was co-created by Memorial Sloan Kettering Cancer Center and multiple global clinical centers. It contains 33,126 dermoscopic images from 2,056 patients, with an average of 16 lesion images per patient. The dataset development process involved querying and filtering from multiple clinical image databases to ensure the accuracy of both image quality and diagnostic labels. This dataset is primarily used to support machine learning challenges, particularly those focused on melanoma evaluation and diagnosis, with the goal of improving the diagnostic accuracy and efficiency of melanoma by providing patient-level contextual information.

搜集汇总

背景与挑战

背景概述

该数据集是一个用于黑色素瘤分类挑战的皮肤镜图像集合，包含来自2,056名患者的33,126张图像，平均每位患者有16个病变图像。它由全球多个中心合作创建，经过严格筛选以确保质量和诊断准确性，主要用于支持机器学习评估，旨在通过患者级别的上下文信息提升黑色素瘤的诊断效率和准确性。