HAM10000
收藏资源简介:
HAM10000数据集是由维也纳医科大学皮肤科ViDIR组和昆士兰大学医学院合作创建的大型皮肤病理图像数据集,包含10015张来自不同来源的皮肤色素病变图像。数据集通过多种采集和清洗方法,利用专门训练的神经网络进行半自动化处理。该数据集主要用于学术机器学习训练,旨在通过机器学习技术提高皮肤色素病变的自动诊断能力,并与人类专家的诊断进行比较。数据集中的病例涵盖了所有重要的诊断类别,超过50%的病变已通过病理学确认。
The HAM10000 dataset is a large-scale skin lesion image dataset co-created by the ViDIR Group of the Department of Dermatology, Medical University of Vienna and the School of Medicine, University of Queensland. It contains 10,015 skin pigmented lesion images sourced from various origins. The dataset was processed semi-automatically using a specially trained neural network following multiple collection and cleaning procedures. Primarily intended for academic machine learning training, this dataset aims to enhance the automatic diagnostic performance of skin pigmented lesions via machine learning technologies, and to compare model outputs with diagnoses made by human experts. The cases included in the dataset cover all major diagnostic categories, with over 50% of the lesions confirmed via histopathology.

- 1The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions维也纳医科大学皮肤科ViDIR组 · 2018年



