Ari Cilingiroglu

Development of an AI model for automatic segmentation of ovarian follicles from 3D MRI images.

Methodology: V-Cycle / Complete AI Pipeline

Missions:

- Analysis, preparation, and structuring of 3D MRI volumes

- Definition of medical issues and imaging constraints

- Development of segmentation models (U-Net 2D/3D, CNN)

- Exploratory study with YOLO for preliminary detection

- Medical image processing: filtering, normalization, interpolation, resampling

- Data augmentation (flip, rotation, scaling, etc.)

- Implementation of a complete pipeline: preprocessing -> training -> post-processing

- Creation of a simple user interface to visualize results

- Collaboration with doctors to adjust clinical needs

- Experimental validation and model comparison via cross-validation

- Full autonomy on the AI workflow (data, models, experiments)

Tools & Technologies:

Python | TensorFlow | Keras | PyTorch | OpenCV | NumPy | Pandas | NiBabel | SimpleITK | MONAI | Scikit-Image | Git | Linux | Windows | Jupyter

During my observation internship at the Austrian Hospital Saint-Georges, I discovered the workings of a medical environment by following healthcare professionals in various departments. I observed surgical procedures, laboratory analyses, and patient care processes, which allowed me to better understand hospital organization and medical technologies. Although I did not perform direct medical tasks, this experience provided me with valuable insights into the challenges of the healthcare sector, thereby strengthening my analytical skills and adaptability in critical environments.