Project detail

CT-based Synthetic Data Generation for Advanced Automatic Defect Detection in X-ray Radiography

Duration: 1.3.2025 — 28.2.2026

Funding resources

Vysoké učení technické v Brně - Vnitřní projekty VUT

On the project

This project focuses on developing a novel method for synthetic training data generation to enhance automatic defect detection in X-ray radiography. By leveraging computed tomography, realistic 2D X-ray projection images will be simulated. This approach ensures a diverse and well-anotated dataset. The resulting dataset will be used to train a deep learning model, improving its robustness and accuracy. This method aims to advance AI-driven X-ray inspection, reducing the need for large real-world datasets while improving its performance.

Mark

CEITEC VUT-J-25-8900

Default language

Czech

People responsible

Tkadlecová Markéta, Ing. - principal person responsible
Zikmund Tomáš, doc. Ing., Ph.D. - fellow researcher

Units

Advanced instrumentation and methods for material characterization
- responsible department (3.3.2025 - not assigned)
Central European Institute of Technology BUT
- responsible department (31.1.2025 - 3.3.2025)
Advanced instrumentation and methods for material characterization
- beneficiary (1.3.2025 - 28.2.2026)

Responsibility: Tkadlecová Markéta, Ing.