Research Library

This study evaluates computer-vision models for automatically identifying vehicle damage from photographs in an insurance setting. The researchers compare ensembles based on YOLOv5 and YOLOv8, focusing on detection accuracy, processing speed, and reducing false…

This paper describes Insoore AI, a claims-management platform that uses photographs to identify vehicle panels, classify damage such as dents and scratches, estimate severity, and support repair-versus-replace decisions. It also discusses the practical difficulty…

This research investigates whether a diffusion-based object-detection model can improve automatic recognition and classification of vehicle damage. It builds on earlier insurance-claims research and examines whether newer AI architectures can better identify subtle or…