marmal88/skin_cancer
收藏资源简介:
--- dataset_info: features: - name: image dtype: image - name: image_id dtype: string - name: lesion_id dtype: string - name: dx dtype: string - name: dx_type dtype: string - name: age dtype: float64 - name: sex dtype: string - name: localization dtype: string splits: - name: train num_bytes: 2490501038.358 num_examples: 9577 - name: test num_bytes: 351507473.24 num_examples: 1285 - name: validation num_bytes: 681758880.144 num_examples: 2492 download_size: 3693626934 dataset_size: 3523767391.7419996 task_categories: - image-classification - image-segmentation language: - en tags: - skin_cancer - HAM10000 pretty_name: HAM10000 size_categories: - 1K<n<10K --- # The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions - Original Paper and Dataset [here](https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/DBW86T) - Kaggle dataset [here](https://www.kaggle.com/datasets/kmader/skin-cancer-mnist-ham10000?resource=download) # Introduction to datasets Training of neural networks for automated diagnosis of pigmented skin lesions is hampered by the small size and lack of diversity of available dataset of dermatoscopic images. We tackle this problem by releasing the HAM10000 ("Human Against Machine with 10000 training images") dataset. We collected dermatoscopic images from different populations, acquired and stored by different modalities. The final dataset consists of 10015 dermatoscopic images which can serve as a training set for academic machine learning purposes. Cases include a representative collection of all important diagnostic categories in the realm of pigmented lesions: Actinic keratoses and intraepithelial carcinoma / Bowen's disease (akiec), basal cell carcinoma (bcc), benign keratosis-like lesions (solar lentigines / seborrheic keratoses and lichen-planus like keratoses, bkl), dermatofibroma (df), melanoma (mel), melanocytic nevi (nv) and vascular lesions (angiomas, angiokeratomas, pyogenic granulomas and hemorrhage, vasc). More than 50% of lesions are confirmed through histopathology (histo), the ground truth for the rest of the cases is either follow-up examination (follow_up), expert consensus (consensus), or confirmation by in-vivo confocal microscopy (confocal). The test set is not public, but the evaluation server remains running (see the challenge website). Any publications written using the HAM10000 data should be evaluated on the official test set hosted there, so that methods can be fairly compared. - Test site can be accessed [here](https://challenge.isic-archive.com/landing/2018/) # Disclaimer and additional information This is a contribution to open sourced data in hugging face for image data. Images can be obtained from above links. Train test split was done using a stratified splitting by cancer/diagnosis type. The code to stratify the dataset can be obtained on my github [here](https://github.com/marmal88/skin_cancer). I do not own any rights to above images. [More Information needed](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
数据集信息: 特征: - 名称:image(图像),数据类型:图像 - 名称:image_id(图像ID),数据类型:字符串 - 名称:lesion_id(皮损ID),数据类型:字符串 - 名称:dx(诊断结果),数据类型:字符串 - 名称:dx_type(诊断类型),数据类型:字符串 - 名称:age(年龄),数据类型:64位浮点型 - 名称:sex(性别),数据类型:字符串 - 名称:localization(皮损部位),数据类型:字符串 数据集划分: - 名称:训练集(train),字节数:2490501038.358,样本数:9577 - 名称:测试集(test),字节数:351507473.24,样本数:1285 - 名称:验证集(validation),字节数:681758880.144,样本数:2492 下载大小:3693626934 字节 数据集总大小:3523767391.7419996 字节 任务类别: - 图像分类 - 图像分割 语言:英语 标签: - 皮肤癌 - HAM10000 友好名称:HAM10000 样本规模类别:1K < n < 10K # HAM10000数据集:常见色素性皮损多来源皮肤镜图像大型集合 - 原始论文与数据集:[此处](https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/DBW86T) - Kaggle数据集:[此处](https://www.kaggle.com/datasets/kmader/skin-cancer-mnist-ham10000?resource=download) ## 数据集简介 用于色素性皮损自动诊断的神经网络训练,常受制于现有皮肤镜图像数据集规模偏小且多样性不足的问题。为此我们发布HAM10000(Human Against Machine with 10000 training images,即“对抗机器的人类:10000张训练图像”)数据集。我们从不同人群中采集皮肤镜图像,并通过不同模态完成图像的获取与存储。最终数据集包含10015张皮肤镜图像,可用于学术场景下的机器学习训练。 该数据集涵盖色素性皮损领域全部重要诊断类别的代表性样本:光化性角化病与上皮内癌/鲍恩病(akiec)、基底细胞癌(bcc)、良性角化样皮损(日光性雀斑/脂溢性角化病及苔藓样角化病,bkl)、皮肤纤维瘤(df)、黑色素瘤(mel)、黑素细胞痣(nv)以及血管性皮损(血管瘤、角化性血管瘤、化脓性肉芽肿及出血,vasc)。 超过50%的皮损经组织病理学(histo)确认,其余病例的真实标签分别为随访检查结果(follow_up)、专家共识(consensus)或活体共聚焦显微镜检查确认(confocal)。 测试集未对外开放,但评估服务器仍在运行(详见挑战赛官网)。任何使用HAM10000数据集撰写的学术成果,均需在该平台提供的官方测试集上开展评估,以确保不同方法间的公平比较。 - 测试站点可通过[此处](https://challenge.isic-archive.com/landing/2018/)访问 ## 免责声明与补充信息 本数据集是针对Hugging Face开源图像数据的贡献。图像可通过上述链接获取。 训练集与测试集的划分采用按癌症/诊断类型分层的策略。用于实现分层划分的代码可在我的GitHub仓库[此处](https://github.com/marmal88/skin_cancer)获取。 本人不拥有上述图像的任何版权。 [需补充更多信息](https://github.com/huggingface/datasets/blob/main/CONTRIBUTING.md#how-to-contribute-to-the-dataset-cards)
数据集概述
数据集名称
- 名称: HAM10000
- 别名: Human Against Machine with 10000 training images
数据集特征
- 图像: 类型为image
- 图像ID: 类型为string
- 病变ID: 类型为string
- 诊断: 类型为string
- 诊断类型: 类型为string
- 年龄: 类型为float64
- 性别: 类型为string
- 局部化: 类型为string
数据集分割
- 训练集: 9577个样本,占用2490501038.358字节
- 测试集: 1285个样本,占用351507473.24字节
- 验证集: 2492个样本,占用681758880.144字节
数据集大小
- 下载大小: 3693626934字节
- 数据集大小: 3523767391.7419996字节
任务类别
- 图像分类
- 图像分割
语言
- 英语
标签
- 皮肤癌
- HAM10000
美观名称
- HAM10000
大小类别
- 1K<n<10K
数据集描述
- 包含: 10015个皮肤病学图像,涵盖多种色素性皮肤病变类型
- 用途: 用于学术机器学习的训练集
- 病变类型: Actinic keratoses and intraepithelial carcinoma / Bowens disease (akiec), basal cell carcinoma (bcc), benign keratosis-like lesions (solar lentigines / seborrheic keratoses and lichen-planus like keratoses, bkl), dermatofibroma (df), melanoma (mel), melanocytic nevi (nv) and vascular lesions (angiomas, angiokeratomas, pyogenic granulomas and hemorrhage, vasc)
- 验证方式: 超过50%的病变通过组织病理学确认,其余通过随访检查、专家共识或体内共聚焦显微镜确认
- 测试集: 不公开,但评估服务器持续运行




