DoctOral-AI

该数据集名为DoctOral-AI，包含535张不同尺寸的图像，用于多类别的口腔病变检测任务，病变类别包括肿瘤性、阿弗他性和创伤性。数据集按照70%用于开发、30%用于测试的比例进行划分，其中开发集进一步细分为80%用于训练、20%用于验证。所有图像都被调整为800x800的尺寸。该数据集的规模为535张图像，所承担的任务是口腔病变检测。The dataset named DoctOral-AI contains 535 images of varying dimensions for multi-class oral lesion detection tasks. The lesion categories include neoplastic, aphthous and traumatic types. The dataset is split into 70% for development and 30% for testing, and the development set is further subdivided into 80% for training and 20% for validation. All images have been resized to 800×800 pixels. This dataset comprises 535 images and its targeted task is oral lesion detection.

搜集汇总

背景与挑战

背景概述

DoctOral-AI是一个专注于口腔癌诊断的数据集，利用可解释人工智能技术，由UniPa和UniPi的多个部门联合研究开发。该数据集可能包含病例数据、注释工具、演示、源代码以及训练和测试报告，旨在支持口腔癌的AI辅助诊断研究。