Artificial Intelligence Research Group
A multidisciplinary research and development unit aimed at developing AI-assisted diagnostic and decision support algorithms for the analysis of medical imaging data
Medical imaging can be thought of as the eye of modern medicine. Just like a detective examining clues at a crime scene, radiologists and other specialists search for traces of diseases through images.
This process is both extremely sensitive and time-consuming area of expertise. At this point, artificial intelligence comes into play, taking on the role of a powerful assistant that can detect details the human eye cannot see and analyze millions of images within seconds.
Our group's main goal is to make imaging-based diagnostic processes faster, more sensitive, and more accessible. This process is not just a technological development, but also means the democratization of healthcare services.
Main application areas where we develop AI-assisted solutions in medical imaging
Automatic identification of anatomical structures and precise segmentation algorithms
Early diagnosis algorithms with classification and malignancy assessment
Integrated analysis of CT, MR, PET, ultrasound, X-ray images
AI-assisted disease progression analysis and risk assessment
Quantitative image characterization and tissue analysis
Decision justification and transparent algorithm development
Our AI-assisted diagnostic systems used and tested in real clinical environments
Detects nodules of 3 millimeters and above with high sensitivity in low-dose computed tomography images and automatically reports growth rate.
Platform that automatically segments multiple sclerosis lesions from T2 and FLAIR images, performs volume measurement, and detects new lesions.
System that classifies coronary artery stenosis with high accuracy, calculates calcium score, and provides cardiac risk stratification.
System that performs diabetic retinopathy staging by detecting microaneurysms, exudates, and neovascularization from fundus photographs.
With our multidisciplinary approach, we collaborate with experts from health and engineering fields
Main research topics we focus on for the development of artificial intelligence technologies in medical imaging
We develop algorithms that try to mimic the human brain's ability to distinguish anatomical structures. We create sophisticated systems that answer the question "which organ does this belong to?" for each pixel.
We combine information obtained from different imaging methods such as CT, MR, PET to reveal the complete picture of the disease. This process is like different musical instruments coming together in a symphony.
We extract quantitative features from images that cannot be selected by the human eye. Tumor heterogeneity, symmetry, surface roughness and other characteristics are characterized with mathematical parameters.
For doctors to trust the algorithm's decisions, we present the justifications for these decisions in an understandable way. It is critical that the system can explain not only the correct answer but also how it reached this answer.
We research hybrid working models where artificial intelligence will support human experts. Like modern aircraft piloting: automatic systems take control, the pilot makes critical decisions.
We develop systems that can perform instant analysis while images are being taken for surgical applications and emergency situations. Like a sound technician at a live concert: the ability to make instant adjustments while the music is playing.
Concrete achievements and clinical impacts of our Research Group in the field of medical imaging
Would you like to collaborate with us to develop artificial intelligence applications in the field of medical imaging?
System development with hospitals and clinics using real data
Inter-university research projects and joint publications
Product development and commercialization with technology companies