MOPG-7: A Multi-Clinic Dental Panoramic Radiograph Dataset with Expert YOLO Labels

MOPG-7 is a publicly available, multi-clinic dental imaging database that aims at furthering studies on artificial intelligence (AI), computer vision, and computer-aided diagnosis (CAD) based on panoramic dental radiographs (orthopantomogram or OPG). This database includes 2,095 completely anonymized panoramic dental radiographs gathered retrospectively from four separate dental centers in Bangladesh such as Sonia Nursing Home, Tangail (1,539 radiographs); Ibn Sina Diagnostic & Consultation Center, Dhaka (533 radiographs); Niramoy Diagnostic Center, Tangail (114 radiographs); and Health City Diagnostic Center, Gaibandha (15 radiographs). Every image comes with a bounding box annotation file in YOLO format (.txt) that has been validated by experts, making it easier for use with modern object detection models like Ultralytics YOLO. The initial bounding boxes were created by a licensed dentist and reviewed independently by another experienced dental professional. A rigorous process of quality control followed after that to remove 127 bounding boxes that were not up to the mark, resulting in 9,834 bounding boxes being used. Dataset Classes The dataset includes annotations for seven clinically relevant dental categories: Missing Teeth: 2,609 annotations Dental Crown: 1,984 annotations Root Canal: 1,956 annotations Caries: 1,410 annotations Wisdom Teeth: 869 annotations Broken Down Teeth: 795 annotations Healthy Teeth: 211 annotations Dataset Contents The released dataset includes: Panoramic dental radiographs (.png) YOLO bounding-box annotation files (.txt) Class definition file (classes.txt) Documentation (README.md) describing the dataset structure, annotation format, and usage instructions Potential Research Applications MOPG-7 was built to enable a broad range of uses for research and learning purposes such as multi-class dental object detection, localization of dental abnormalities, CAD, deep learning for medical imaging, computer vision research, transfer learning and building foundation models, XAI, medical image analysis, object detection algorithm benchmarking, AI-enabled dental diagnosis, dental AI learning, and reproducibility research. Benchmark Performance In order to have an effective baseline, YOLOv11m was trained and tested on MOPG-7 to give a precision score of 90.2%, recall score of 91.5%, mAP@0.5 of 93.0%, and mAP@0.5:0.95 of 59.2%. Of the seven classes, the Wisdom Teeth and Dental Crown were those that performed best, while Missing Teeth and Caries posed more difficulties because of anatomical differences.MOPG-7是一款公开可用的多诊所牙科影像数据库，旨在推动基于全景牙科X光片（orthopantomogram，简称OPG）的人工智能（Artificial Intelligence，AI）、计算机视觉以及计算机辅助诊断（computer-aided diagnosis，CAD）相关研究。该数据库回溯性收集了来自孟加拉国四家独立牙科中心的2095张完全匿名的全景牙科X光片：坦盖尔市索尼娅疗养院（1539张）、达卡市伊布·西纳诊断与咨询中心（533张）、坦盖尔市尼拉莫伊诊断中心（114张）以及盖班达市健康城诊断中心（15张）。 每张影像均附带YOLO格式的边界框标注文件（.txt），且经专家审核验证，可便捷适配Ultralytics YOLO等现代目标检测模型。初始边界框由执业牙医绘制，并由另一名经验丰富的牙科专业人员独立复核；随后通过严格的质量控制流程，剔除了127个不合格标注框，最终可用边界框共计9834个。 数据集类别 本数据集包含7类临床相关的牙科标注类别： 缺牙（Missing Teeth）：2609个标注 牙冠（Dental Crown）：1984个标注 根管治疗（Root Canal）：1956个标注 龋病（Caries）：1410个标注 智齿（Wisdom Teeth）：869个标注 残损牙体（Broken Down Teeth）：795个标注 健康牙齿（Healthy Teeth）：211个标注 数据集内容 本次发布的数据集包含以下内容： 全景牙科X光片（.png格式） YOLO边界框标注文件（.txt格式） 类别定义文件（classes.txt） 用于说明数据集结构、标注格式与使用指南的文档（README.md） 潜在研究应用场景 MOPG-7的构建旨在支撑广泛的研究与学习用途，涵盖多类别牙科目标检测、牙科异常定位、计算机辅助诊断、医学影像深度学习、计算机视觉研究、迁移学习与基础模型构建、可解释人工智能（eXplainable AI，XAI）、医学影像分析、目标检测算法基准测试、AI辅助牙科诊断、牙科人工智能学习以及可重复性研究。 基准测试性能 为建立有效基线模型，研究人员在MOPG-7数据集上训练并测试了YOLOv11m模型，其精准率达90.2%、召回率达91.5%、mAP@0.5为93.0%、mAP@0.5:0.95为59.2%。在7个标注类别中，智齿与牙冠的检测性能最优，而缺牙与龋病则因解剖结构差异，成为检测难点。