TMC-Tongue: A standardized tongue image dataset with pathological annotations for AI-assisted TCM diagnosis

This dataset contains 21 disease categories that can be used for target detection in tongue diagnosis. The categories are jiankangshe (Healthy Tongue), botaishe (Tongue with peeling coating), hongshe (Red tongue), zishe (Purple tongue), pangdashe (Chubby tongue), shoushe (Thin tongue), hongdianshe (Red dot tongue), liewenshe (Cracked tongue), chihenshe (Dentate tongue), baitaishe (White coating tongue), huangtaishe (Yellow coating tongue), heitaishe (Black coating tongue), huataishe (Smooth coating tongue), shenquao (renal depression), shenqutu (renal protrusion), gandanao (Hepatobiliary depression), gandantu (Hepatobiliary protrusion), piweiao (spleen and stomach depression), xinfeitu (heart and lung protrusion), xinfeiao (heart and lung depression), corresponding numerical order is 0-19.
Among them, there are 5594 images in the training set, 572 images in the validation set, and 553 images in the test set. Contains three annotation formats: coco/. txt/. XML, which can be used for expe..., , # TMC-Tongue: A standardized tongue image dataset with pathological annotations for AI-assisted TCM diagnosis

Dataset DOI: [10.5061/dryad.1c59zw48r](10.5061/dryad.1c59zw48r)

## Description of the data and file structure

Using high-precision cameras to capture volunteers' tongues to obtain data, all data only includes volunteers' tongues and mouths, and does not involve volunteers' facial information. The dataset only contains photos and feature labels, and does not contain any personal privacy information of volunteers.

### Files and variables

The dataset consists of three folders, namely shezhenv3 coco, shezhenv3 txt, and shezhenv3 xml, which can be used by three different object detection models. The image data in the three folders is the same, but the label files are not the same. Users can choose the file that suits them according to their own needs. These three folders all contain test sets, training sets, and validation sets, each containing image data and corresponding label...,

本数据集包含21类可用于舌诊目标检测的病症类别，具体类别依次为：健康舌（Healthy Tongue）、舌苔剥落舌（Tongue with peeling coating）、红舌（Red tongue）、紫舌（Purple tongue）、胖大舌（Chubby tongue）、瘦舌（Thin tongue）、红点舌（Red dot tongue）、裂纹舌（Cracked tongue）、齿痕舌（Dentate tongue）、白苔舌（White coating tongue）、黄苔舌（Yellow coating tongue）、黑苔舌（Black coating tongue）、光滑苔舌（Smooth coating tongue）、肾凹（renal depression）、肾凸（renal protrusion）、肝胆凹（Hepatobiliary depression）、肝胆凸（Hepatobiliary protrusion）、脾胃凹（spleen and stomach depression）、心肺凸（heart and lung protrusion）、心肺凹（heart and lung depression），对应的数字编号为0至19。
其中训练集包含5594张图像，验证集包含572张图像，测试集包含553张图像。本数据集支持三种标注格式：COCO、.txt、.XML，可用于相关实验……

# TMC-Tongue：面向AI辅助中医诊断的带病理标注标准化舌图像数据集
数据集DOI：10.5061/dryad.1c59zw48r

## 数据与文件结构说明
采用高精度相机采集志愿者舌部图像以构建数据集，所有数据仅包含志愿者舌部与口腔区域，未涉及面部信息。本数据集仅包含图像与特征标签，未包含志愿者任何个人隐私信息。

### 文件与变量说明
本数据集包含三个文件夹，分别为shezhenv3 coco、shezhenv3 txt与shezhenv3 xml，可适配三种不同的目标检测模型。三个文件夹内的图像数据完全一致，但标签文件格式存在差异，用户可根据自身需求选择适配的文件格式。三个文件夹均包含测试集、训练集与验证集，每个子集均包含图像数据与对应标签……