"Pixemantic Vehicle Damage-29"

"The Pixemantic Vehicle Damage-29 dataset, developed by Pixemantic, is designed for instance segmentation and object detection of vehicles, focusing on both car components and damage types. It contains 29 classes, covering structural parts (e.g., doors, bumpers, wheels) and damage indicators (e.g., scratches, dents, cracks, corrosion, broken parts). The dataset includes 3,850 training images, 930 validation images, and 464 test images, with annotations suitable for bounding boxes and segmentation masks.This dataset was developed to support the creation of digital sSinistres for fraud detection in insurance and smart automotive auditing systems. It is compatible with state-of-the-art instance segmentation and detection models such as:Roboflow-3-n-segYOLOv11YOLOv11n-segVehicleDamage-29 provides a comprehensive benchmark for deep learning models in real-world automotive scenarios, enabling accurate damage detection, vehicle inspection, and insurance fraud prevention."