Multi-Source Dental X-Ray Dataset Using Image-to-Image Transformation
收藏资源简介：
The Teeth View X-ray Image Dataset is a collection of dental X-ray images gathered from different dental clinics. It is designed for machine learning tasks such as object detection. The dataset is organized into one main folder: the object detection dataset. The object detection folder contains 1,674 augmented images with corresponding labels in JSON format. By applying a diffusion model image to image, we have generated additional X-ray images of teeth scans to enhance the Teeth View X-ray Image Dataset to be deeper, more diverse, and beneficial for model tuning. The synthetic image contains various anatomy, density, and acquiring conditions of the tooth and thereby enhances the model's robustness for unusual dental conditions. The increased size of the dataset works to counterbalance class imbalance by having more examples of the under-represented classes and reducing model bias. Also, diffusion models demonstrate high effectiveness in reproducing noise patterns, making training more robust; improving the clarity of images through denoising processes; and enabling highly precise segmentation mask predictions to facilitate efficient boundary detection. Variables: BDC-BDR Teeth -208 Caries Teeth - 342 Fractured Teeth -52 Healthy Teeth - 893 Impacted Teeth - 348 Inflection Teeth - 92 Folder structure: Teeth view X-ray Image dataset | |---Objectect Detection | |----image |----Labels.
牙齿视角X光图像数据集（Teeth View X-ray Image Dataset）是从多家牙科诊所采集的牙科X光影像集合，专为目标检测（object detection）等机器学习任务构建。 该数据集仅设置一个主文件夹：目标检测数据集。目标检测文件夹内包含1674张经过数据增强的影像，以及配套的JSON格式标注文件。 通过应用图像到图像的扩散模型（diffusion model），我们生成了额外的牙齿扫描X光影像，以扩充该数据集的规模与维度，提升其多样性，使其更适配模型调优需求。合成影像涵盖了牙齿的多种解剖结构、密度与成像条件，可有效提升模型针对罕见牙科病症的鲁棒性。 扩充后的数据集规模通过为代表性不足的类别提供更多训练样本，有效缓解了类别不平衡问题，同时降低了模型的偏差。此外，扩散模型在重现噪声模式方面表现优异，可使训练过程更具鲁棒性；通过降噪流程提升影像清晰度；还能生成高精度的分割掩码（segmentation mask）预测结果，助力高效的边界检测任务。 类别样本统计： BDC-BDR牙齿 - 208 龋病牙齿（Caries Teeth） - 342 折裂牙齿（Fractured Teeth） - 52 健康牙齿（Healthy Teeth） - 893 阻生牙齿（Impacted Teeth） - 348 形态异常牙齿（Inflection Teeth） - 92 文件夹结构： 牙齿视角X光图像数据集 | ├── 目标检测（Object Detection） │ │ ├── image │ └── Labels.
