About Me
I am a PhD Candidate in Computer Vision at Julius-Maximilians-Universität Würzburg (expected graduation: December 2026) and a Research Associate & Software Engineer at the Interactive Media Technology Lab at Technische Hochschule Nürnberg Georg Simon Ohm.
My work bridges fundamental visual computing research and high-performance software engineering. In my doctoral thesis, I investigate “Multi-view Amodal Instance Segmentation for 3D Scene Representation and Novel View Synthesis.” Across our research lab, I have led and contributed to computer vision, visual computing, and machine learning software projects in C++ and Python with a strict engineering focus on execution speed, GPU parallelization, and numerical accuracy—resulting in peer-reviewed publications at CVPR 2024, Eurographics 2024, WSCG 2025, VRIH, ICIPCW 2023, and VISAPP 2026.
Alongside academic research, I gained industrial software engineering experience at mgm technology partners, where I developed and maintained production frontend web applications for major retail chains in TypeScript and JavaScript and built embedded text messaging prototypes in Python and TypeScript. I also assist in teaching undergraduate Software Engineering and Object-Oriented Programming and serve as a peer reviewer for the 3D Modeling track at ECMS 2025 & 2026.
Research & Engineering Interests
- 3D Scene Representation & Gaussian Splatting
- Multi-View Amodal Instance Segmentation
- Dense Optical Flow & Correspondence Estimation
- High-Performance C++ & GPU Optimization
- Foundation Models (SAM3) & Video Segmentation
- Software Architecture & Frontend Web Engineering
Education
-
PhD in Computer VisionJulius-Maximilians-Universität Würzburg • Nov 2024 – Present (Exp. Dec 2026)Thesis: “Multi-view Amodal Instance Segmentation for 3D Scene Representation and Novel View Synthesis”
-
M.Sc. in Media Informatics (Grade: 1.0)Technische Hochschule Nürnberg Georg Simon Ohm • Mar 2021 – Dec 2022
-
B.Sc. in Computer Science (Grade: 1.4)Technische Hochschule Nürnberg Georg Simon Ohm • Oct 2017 – Mar 2021
Publications
Software Engineering & Research Highlights
C++ • Python • TypeScriptAuto-Terminating Hierarchical Histogram Oversegmentation
Engineered a novel hierarchical histogram threshold segmentation algorithm in C++ that automatically adjusts to local image detail and self-terminates without manual superpixel tuning.
Dense 3D Gaussian Splatting Initialization
Co-developed a fast, dense point-cloud initialization pipeline in C++ and Python designed for sparse multi-view image inputs, improving novel view synthesis convergence and visual fidelity.
EGO-VC: Evolutionary GPU Correspondence Optimization
Supervised and architected a parallel evolutionary optimization pipeline on the GPU in C++ for rapid, high-precision visual patch alignment across views.
MH-Flow & Promptable SAM3 Adaptation for STEP
Developed multi-homography optical flow estimation for occluded scenes (C++/Python) and co-developed a hybrid Python framework adapting the promptable SAM3 foundation model to STEP.
Multidimensional VR Image Morphing & DFM4SFM
Contributed high-speed C++ implementations for real-time image-based rendering of open 3D/VR environments and dense feature matching for accurate Structure-from-Motion point clouds.
Enterprise Retail Web Applications & Messaging Prototypes
Developed and maintained production frontend web applications for major retail chains in TypeScript/JavaScript and built embedded text messaging software prototypes in Python & TypeScript.
Experience
Research Associate
Led research and software projects in computer vision, visual computing, and machine learning (C++, Python):
- Developed, published, and presented a novel hierarchical histogram threshold segmentation method for auto-adjusting high-detail oversegmentation (C++, focus on accuracy; CVPR 2024).
- Co-developed and published a dense 3D Gaussian Splatting initialization for sparse image data and better visual quality (C++, Python, focus on speed and accuracy; Eurographics 2024).
- Supervised, published, and presented an evolutionary GPU-optimization of visual correspondences (EGO-VC) for aligning image patches (C++, focus on speed; WSCG 2025).
- Contributed to research and development of dense feature matching for Structure from Motion (DFM4SFM) for dense, accurate point cloud reconstruction and robust camera pose estimation (C++, focus on speed; ICIPCW 2023).
- Contributed to research and development on multi-view optical flow estimation (MH-Flow) with focus on occlusion areas (C++, Python, focus on accuracy; VISAPP 2026).
- Co-developed and submitted a hybrid framework that adapts the promptable SAM3 foundation model to the STEP task (Python, focus on accuracy; Submitted to WACV 2027).
- Peer reviewed 3D Mod track at ECMS 2025 and ECMS 2026.
Student Research Assistant
Assisted with research and development for lab projects in C++ and Python. Contributed to the development, evaluation, and publishing of fast image-based rendering of open 3D and VR environments (C++, focus on speed; published in VRIH journal 2023/2025).
Working Student — Software Development (Frontend)
- Developed and maintained web applications for large retail chains (TypeScript, JavaScript).
- Gained practical experience in a professional software development environment and working in larger projects (Git, Jira).
Intern — Software Development (Frontend)
- Developed experimental software application prototypes for embedded text messaging (Python, TypeScript).
- Gained practical experience in a professional software development environment (Git).
Teaching & Academic Service
Object-Oriented Programming
Assisted teaching at TH Nürnberg covering object-oriented design principles, encapsulation, inheritance, polymorphism, and clean coding practices.
Software Engineering
Assisted teaching at TH Nürnberg guiding students through software architecture, design patterns, testing, Git workflows, and collaborative project engineering.
Reviewing & Networks
- • Peer Reviewer: 3D Modeling (3D Mod) Track at ECMS 2025 & ECMS 2026.
- • Member: BayWISS Verbundkolleg Digitalisierung (Nov 2024 – Oct 2027).
Technical Skills & Languages
Programming
Domains & Methods
Computer Vision, Image Processing, Machine Learning, Data Mining, Frontend Web Development, Mobile App Development, Prototyping, Academic Publishing, Peer Reviewing
Frameworks & Tools
Languages
- German Native
- English Fluent / Excellent Written