Publications & Patents
Scientific publications, patents and research reports from our team.
Patents and Intellectual Property Rights
During his work as a scientist and project manager at Technische Hochschule Ingolstadt, Dr. Mühenad Bilal was involved as inventor in several national and international intellectual property applications. The developments were created in the fields of optical measurement technology, innovative illumination systems and AI-based inspection methods.
EP4130720 – Object Illumination
| Status | European patent granted |
| Role | Inventor |
| Patent holder | Technische Hochschule Ingolstadt |
WO2025/021628 – Sample Illumination
| Status | International publication |
| Role | Inventor |
| Applicant | Technische Hochschule Ingolstadt |
PCT Application
PCT/EP2024/070322 – Sample Illumination
| Status | International patent application |
| Role | Inventor |
| Applicant | Technische Hochschule Ingolstadt |
The patents and patent applications listed were developed during Dr. Mühenad Bilal's work at Technische Hochschule Ingolstadt. The intellectual property rights belong to the respective patent holder. This presentation serves exclusively to document scientific achievements and involvement as inventor.
Scientific Publications
Peer-reviewed articles, conference papers and research reports.
Automatized End Mill Wear Inspection Using a Novel Illumination Unit and Convolutional Neural Network
Mühenad Bilal, Ranadheer Podishetti, Leonid Koval, Mahmoud A. Gaafar, Daniel Grossmann, Markus Bregulla
A novel method for automated wear detection on end mills – combining a specially designed illumination system with a CNN model. Validation achieved an mIoU value of 0.99 for tool segmentation.
Benchmarking CNN Architectures for Tool Classification: Evaluating CNN Performance on a Unique Dataset Generated by Novel Image Acquisition System
Mühenad Bilal, Ranadheer Podishetti, Daniel Grossmann, Markus Bregulla
Presentation of the ToolSurface-144 dataset with 144 tool classes under varying illumination conditions. Benchmarking of 6 CNN architectures – EfficientNet-B0 achieved a classification accuracy of 99% with fine-tuning.
Interested in Our Research?
Contact us – we welcome scientific exchange and cooperations.
Get in Touch