cv
Basics
| Name | Mertcan Özdemir, PhD |
| Label | Biomedical Engineer · Medical Image Processing · Deep Learning for MRI |
| mertcanozdemir@yahoo.com | |
| Phone | +90 535 355 6389 |
| Url | https://mertcanozdemir.com |
| Summary | Recent PhD graduate in Biomedical Engineering with research expertise in deep learning-based cardiac MRI synthesis using diffusion models and attention mechanisms. Experienced in MR image reconstruction pipelines, DICOM data processing, and generative AI for medical imaging. Seeking to expand into ultra-high field MRI acquisition and reconstruction, combining computational skills with hands-on scanner experience. |
Work
- 2023.11 - 2025.12
Teaching and Research Assistant
TOBB University of Economics and Technology
Department of Biomedical Engineering.
- Taught Biomedical Image Processing (BMM411): frequency-domain filtering, Wiener filters, Radon transform, wavelet transforms, and color image processing.
- Supervised undergraduate thesis projects in medical image analysis and signal processing.
- Coordinated Biomedical Instrumentation Laboratory: ECG, EMG, EEG, and EOG signal acquisition and processing with MATLAB.
- 2020.04 - 2023.11
R&D Engineer (Biomedical)
Okuman Medikal Sistemler A.Ş.
- Led three TÜBİTAK-funded medical device R&D projects (~€500K total budget); managed technical development from concept to prototype.
- Developed embedded software architecture compliant with IEC 62304 for patient monitoring devices.
- Managed IEC 60601-1 (electrical safety) and IEC 60601-1-2 (EMC) compliance testing and documentation.
- 2017.05 - 2020.04
Research Assistant
TOBB University of Economics and Technology
-Assisted the supervisor with research projects, laboratory courses, and documentation in biomedical engineering.
- 2016.01 - 2016.04
- 2015.05 - 2015.08
Education
-
2019.01 - 2025.01 PhD
TOBB University of Economics and Technology
Biomedical Engineering
Thesis: Cardiac MRI Image Data Augmentation Using Diffusion Model-Based Asymmetric Attention Mechanisms
- Developed a novel asymmetric attention-enhanced diffusion model for synthesizing realistic cardiac cine MRI sequences from the OCMR dataset.
- Achieved FID score of 77.78; radiologists distinguished synthetic from real images with only 60% accuracy in clinical validation.
- Extensive experience with k-space data, DICOM pipelines, cardiac cine reconstruction, and evaluation metrics (FID, SSIM, MS-SSIM).
-
2017.01 - 2019.01 MSc
TOBB University of Economics and Technology
Biomedical Engineering
Thesis: Investigation of Ballistic Gelatin-Based Phantom Properties for Ultrasound Training
-
2013.01 - 2017.01
Volunteer
- 2023.01 - Present
- 2021.01 - Present
Recycler
ÇEVKO Çevre Koruma ve Ambalaj Atıkları Değerlendirme Vakfı
Firmamda ambalaj atığı geri dönüşüm projesini başlattım ve yılda yaklaşık 150 kg ambalaj atığının geri dönüşüme kazandırılmasını sağladım.
- 2020.02 - Present
Certificates
| IIHF Hockey Development Camp - Finland | ||
| International Ice Hockey Federation (IIHF) |
| Python Data Structures | ||
| University of Michigan |
| AI For Everyone | ||
| DeepLearning.AI |
| Neural Networks and Deep Learning | ||
| deeplearning.ai |
| Data or Specimens Only Research | ||
| CITI Program |
| C Programming | ||
| Sololearn |
| Introduction to Medical Software | ||
| Yale University |
Publications
-
2025.01.01 Cardiac MRI Image Data Augmentation Using Diffusion Model-Based Asymmetric Attention Mechanisms
Diagnostics (Q1 Journal)
A novel asymmetric attention-enhanced diffusion model for synthesizing realistic cardiac cine MRI sequences from the OCMR dataset. The model achieved an FID score of 77.78, with radiologists distinguishing synthetic from real images with only 60% accuracy in clinical validation.
-
2023.08.31 Thermal Monitoring System (International Patent WO2023163682A1)
WIPO — International Patent
Thermal camera integration for neonatal radiant warmers.
-
2019.11.11 Wireless Transmission of Heart Rate and Blood Pressure Measurements for Remote Patient Monitoring
IEEE Xplore
This paper describes the development of a wireless pulse and blood pressure monitoring system based on a microprocessor with low cost. It is aimed to receive blood pressure information (tension) through certain processes such as amplifying and filtering of the signal which is obtained by using BIOPAC SS19LA Blood Pressure Cuff and Transducer. With the designed prototype, it is possible to determine both systolic and diastolic pressure. Following preprocessing of the signal, the information of blood pressure and heart rate data of the patient is wirelessly transmitted to smartphone by using Bluetooth Module. The wireless system is used to transmit the measured data to a remote location. In this way, the biomedical parameters obtained from the patient can be instantly transmitted to any pre-paired smart phone and also recorded in the corresponding device.
-
2018.05.30 Investigating Ballistic Gelatin Based Phantom Properties for Ultrasound Training
Springer
The simulation has become an important tool for healthcare practitioners who have difficulty in accessing patients to learn ultrasound imaging modes. The ultrasound phantoms are specially designed objects that are used or imagined to evaluate, analyze and adjust the performance of test devices. These phantoms for ultrasonography devices are expensive, and low-cost alternatives have provided an educational experience that does not give the best result. Ballistic gelatin is a member of the 250-Bloom hydrogel family that resembles human muscle tissue in terms of its mechanical properties. The 250-Bloom Ballistic Gelatin (BG) is prepared with different mixing ratios to be made the mechanical tests such as gunshot, compression and electrical conductivity measurement. The results are compared with the mechanical results of human muscle tissue in order to measure the similarity of the 250-Bloom BG we prepared to human muscle tissue. It is showed that the 250-Bloom BG phantom model has very close mechanical properties to human muscle tissue at time-dependent characteristics of mechanical test results. It is also measured how long it can last without degradation with the time required to use it in the simulation and it is coated with the thermal insulation material needed to extend the degradation period. Based on these results, 250-Bloom BG phantom is recommended as a model for the creation of phantom limb model. Consequently, this model is a much more affordable alternative and easy to produce, it facilitates to work with any organ model in ultrasound imaging for healthcare practitioners.
-
2017.10.27 Lower limb phantom design and production for blood flow and pressure tests
The EuroBiotech Journal
Phantoms are specifically designed objects that are utilized or imaged to evaluate, analyze and tune the performance of experimental devices. In this project, it is aimed to design a phantom that responds in a similar manner with how human blood circulation would act in specific flow and pressure tests such as pulse measurement. Ballistic gelatin is a member of hydrogel family with 250 Bloom value which resembles human muscle tissue in terms of mechanical features. That’s why we carried out a uniaxial compression test on our gelatin sample to analyze its similarity of human muscle tissue in terms of elastic modulus, stiffness and rupture strength. Test results indicated that our gelatin sample has approximate values with organic human muscle tissue. Designed model was X-rayed and the similarities of the model to human texture were compared. After producing of lower limb phantoms, we carried out a circulation test through them by the aid of a peristaltic pump to simulate the actual blood circulation of human body limbs. This designed phantom is made ready for available flow and pressure tests.
Skills
| Programming | |
| Python | |
| MATLAB | |
| C | |
| C++ |
| Deep Learning | |
| PyTorch | |
| Diffusion Models | |
| GANs | |
| Attention Mechanisms |
| MRI & Imaging | |
| DICOM | |
| pyDicom | |
| k-space processing | |
| Cardiac cine MRI | |
| Image reconstruction |
| Image & Signal Processing | |
| Frequency-domain filtering | |
| Wavelet transforms | |
| Radon / inverse Radon transform | |
| Digital signal processing |
| Evaluation Metrics | |
| FID | |
| SSIM | |
| MS-SSIM | |
| Clinical validation protocols |
| Tools & Platforms | |
| Git/GitHub | |
| LaTeX | |
| Linux (Arch) | |
| MATLAB Grader | |
| SolidWorks | |
| Simulink |
| Embedded Systems | |
| STM32 | |
| FreeRTOS | |
| Medical software (IEC 62304) |
| Medical Devices | |
| IEC 60601-1 | |
| IEC 60601-1-2 (EMC) | |
| System testing | |
| Project management |
Languages
| Turkish | |
| Native Speaker |
| English | |
| Professional Working Proficiency |
| German | |
| Elementary |
Interests
| Ice Hockey | |
| National Ice Hockey Team member (Turkey) | |
| Founder and captain of the TOBB University ice hockey team |