Kerim Solmaz
Computer vision researcher and PhD candidate. 4 publications.
research
I work on models that let machines see, measure, and say how sure they are.
I am a research assistant and PhD candidate in Computer Engineering at Bilecik Şeyh Edebali University. I started with 3D vision and differentiable rendering in my master's; today I apply the same tools to environmental data.
phd
Reliable water quality prediction with physics-constrained multi-task learning · advised by Prof. Dr. Uğur Yüzgeç
interests
about
- nowPhD · Computer Engineering · Bilecik Şeyh Edebali University
- 2023MSc · Computer Engineering · Iskenderun Technical University
- 2020–23TÜBİTAK project scholar · self-supervised video stabilization
- 2020BSc · Electrical & Electronics Engineering · Iskenderun Technical University
This room is generated in code: every object's depth, surface normal and boundary is known. Try one of the vision modes below.
computer vision
- Video stabilizationSelf-supervised deep learning; 3D point clouds from depth, camera intrinsics/extrinsics and optimisation via differentiable rendering. TÜBİTAK
- Camera localizationComparing optimisation techniques for recovering camera pose from images; relocalization through differentiable rendering. SIU 2022
- Learning-based image renderingA review of rendering with deep networks. JISwA 2022
- Image processing + ML for water treatmentImage features for bacterial cadmium removal. JWPE 2025
publications
- 2026
Efficient removal of Ni, Al, and As from water using Micrococcus yunnanensis: batch experiments and explainable machine learning
A. Alkan Uçkun, Ş. Akkurt, M. K. Solmaz, M. Uçkun
Biodegradation 37(5), 142 · doi
- 2025
Integrating image processing and machine learning for efficient cadmium removal from water by bacteria
M. K. Solmaz, Ş. Akkurt, A. Alkan Uçkun, M. Uçkun
Journal of Water Process Engineering 77, 108475 · doi
- 2022
Performance evaluation of different optimization techniques for image-based camera localization
M. K. Solmaz, M. Sarıgül, L. Karacan
SIU 2022 · IEEE · doi
- 2022
Learning-based image rendering
M. K. Solmaz, M. Sarıgül, L. Karacan
Journal of Intelligent Systems with Applications 5(1) · doi
things i made
- internetprogramlama.onlineA course platform where students write HTML/CSS/JS in the browser and get compared to a target design.
- biodiversityA wiki of Türkiye's species built from GBIF and Wikidata.
- department systemsInternship management and capstone tracking, on Cloudflare.
- this roomModelled in code with three.js; the vase on the printer is reprinted every 90 seconds.
contact
For collaborations, student projects, or just hello.