GulfSCEI is pleased to welcome a new Austrian international research collaborator from Graz University of Technology, Christina Thrainer. A master's student specializing in artificial intelligence and computer vision, Christina will be conducting cutting-edge research on autonomous defect detection in culverts and sewer pipes as part of her master's thesis.

The project, titled AI-Based Culvert-Sewer Inspection, aims to improve traditional defect detection methods by developing an advanced semantic segmentation model. This research will address key challenges such as varying pipe structures, occlusions, and difficult lighting conditions—factors that often hinder existing inspection technologies.

Building on previous research at GulfSCEI, which explored deep learning techniques like U-Net and FPN for defect detection, this project will refine and enhance these approaches. Christina remarks, "over the next five months, I will focus on developing a more robust segmentation model, integrating effective data augmentation strategies, and evaluating performance across multiple datasets. The goal is to improve accuracy, robustness, and generalizability, ensuring superior defect detection in real-world conditions."

Christina is the third visiting graduate research assistant from Graz University of Technology within the last two years. We are excited to continue this collaboration and look forward to working alongside Christina towards AI inspection improvements.