PhD in Computer Science Fellowship
Inverse problems in medical imaging, self-supervised restoration, and uncertainty quantification, with Prof. Müjdat Çetin. Supported by a University of Rochester PhD Fellowship.
İsmail Can Yağmur
PhD Student in Computer Science, University of Rochester
Advised by Prof. Müjdat Çetin
About
I work on uncertainty quantification for machine learning in medical imaging — on making learned models not only accurate, but honest about what they do not know.
My current work applies this to point-of-care diagnostics, in the broader setting of inverse problems and image restoration. I am interested in how such methods behave once they meet real clinical systems rather than benchmarks.
Before Rochester I spent three years on deep-learning research in computer vision — multispectral and cross-spectral image registration, satellite imagery, and signal classification — across several nationally funded research projects.
Research Interests
- Inverse problems
- Computational imaging
- Uncertainty quantification
- Self-supervised learning
- Image registration
- Trustworthy machine learning
Publications
- 2026
- 2026
-
2025
CS-REG-NET: A Self-Supervised Visual-State-Space Based Architecture for Cross-Spectral Registration of Thermal and Optical Imagery SIU 2025First authorBest paper
Best Paper Award at the 33rd Signal Processing and Communications Applications Conference. The accompanying thesis presentation placed 3rd in the conference thesis competition.
- 2025
-
2023
Improved Homographic Adaptation for Keypoint Generation in Cross-Spectral Registration of Thermal and Optical Imagery SPIE Image and Signal Processing for Remote Sensing XXIXFirst author
- 2023
-
2022
Classifying LPI Radar Waveforms with Time-Frequency Transformations Using a Multi-Stage CNN System IEEE International Radar Symposium (IRS)
Education
M.Sc. in Computer Science
Computer vision with Prof. Hasan Fehmi Ateş — deep learning for image matching, and building height estimation from satellite imagery in dense urban areas.
B.Sc. in Computer Science & B.Sc. in Electrical and Electronics Engineering
Ranked first in the Computer Science graduating class; full performance scholarship.
Experience
Software Consultant — Clerio Vision, Inc.
Transforming a manual MATLAB analysis pipeline into a scalable, automated and reproducible data-analysis framework to enable large-scale longitudinal analysis — supporting work on whether optical features can help identify children at high risk of future myopia progression.
Graduate Researcher — Özyeğin University
Research on nationally funded TÜBİTAK projects (Türkiye's NSF equivalent):
- 1001 · 125E200 Adaptive systems for multimodal image matching — 2025
- 1505 · 5230108 Building height estimation and multispectral registration, industry partnership — 2023–2025
- 1003 · 118E891 Wide-area surveillance and cross-spectral registration — 2021–2023
- 2247-C STAR Optical–thermal image matching, undergraduate research fellowship — 2021–2022
Also: LPI radar waveform classification with multi-stage CNNs and time–frequency transforms, and Autonomy Division lead for the OzU Rover Team — a finalist at both the University Rover Challenge and the European Rover Challenge.
AI Software Engineer Intern — Baykar Technologies
Homography applications including video stabilization and compensation.
Teaching
University of Rochester — Teaching Assistant
Özyeğin University — Teaching Assistant, 2023 – 2025
- Introduction to Deep Learning
- Computer Vision
- Computer Networks
- Introduction to Programming
Awards
- PhD Fellowship, University of Rochester
- Best Paper Award, SIU 2025
- 3rd place, SIU 2025 thesis competition
- Graduate full-tuition scholarship, Özyeğin University
- Undergraduate full performance scholarship, Özyeğin University
Languages
- Turkish
- English
- German