Annotated Dental Instrument Dataset for YOLO Object Detection
Description
This dataset contains annotated images of dental surgical instruments designed for object detection and computer vision research. The dataset aims to support the development and evaluation of deep learning models for automatic detection and recognition of dental instruments used in dental and oral surgical procedures. The dataset consists of 847 images covering 22 different types of dental instruments. The instrument categories included are: Aji Bone Cutter, Andrew Tongue Depressor, Angled Artery Forceps, Cross Bar Elevator, Curved Gum Scissors, Dental Aspirating Syringe, Dental Elevator, Frazier Suction Tube, Mandibular Anterior Extraction Forceps, Mandibular Cowhorn Forceps, Mandibular Lower Incisors Forceps, Mandibular Lower Molar Forceps, Maxillary Anterior Canine Forceps, Maxillary Molar Dental Extraction Forceps, Maxillary Upper Incisors Forceps, Maxillary Upper Molar Forceps, Maxillary Upper Molars Extraction Forceps, Mayo Hegar Needle Holder, Orthodontic Ligature Cutter, Straight Artery Forceps, Straight Tweezers, and Wire Orthopedic Cutter. All images were captured using a GoPro camera from a top-view perspective. During data collection, the instruments were placed on white paper sheets on a table surface to create a clean and consistent background suitable for object detection tasks. To enhance dataset diversity and improve model robustness, images were captured under different illumination conditions including: Normal lighting, Bluish lighting conditions, Warm/orange lighting conditions. These variations simulate different real-world imaging environments commonly encountered in clinical or laboratory settings. Additionally, each instrument was photographed from multiple orientations and angles by slightly rotating and repositioning the tools during image acquisition. This approach ensures that the dataset contains diverse perspectives of each instrument, enabling machine learning models to learn more robust visual features. The annotation process was conducted using the Roboflow platform. Each instrument was manually labeled using bounding box annotations following the standard YOLO object detection format. The dataset is compatible with several YOLO-based object detection frameworks, including: YOLOv5 (PyTorch), YOLOv7 (PyTorch), YOLOv8, YOLOv11, This dataset can be used for training, validation, and testing machine learning models in areas such as: computer vision, medical image analysis, intelligent healthcare systems, automated dental instrument detection for robotic surgery. The dataset contains only surgical instruments and no patient data, therefore no ethical approval or patient consent was required.
Files
Institutions
- Manav Rachna International Institute of Research and StudiesHaryana, Faridabad