RESEARCH TEAM

ASSISTANT PROF.DR.
RUKIYE KARAKIS

RESEARCH FIELD: Artificial Intelligence,Bioinformatics, Medical Data Security,Image Processing

ASSISTANT PROF.DR.
KALI GURKAHRAMAN

RESEARCH FIELD: Computer Vision, Image Processing, Artificial Intelligence, Medical Data Security

...

You can send your CV to deepbrain@cumhuriyet.edu.tr for graduate studies

EEG-MRI DATA HIDING TOOL

BACKGROUND:

Epilepsy is a functional disease which becomes an abnormal electrical activity in brain neurons. EEG is used both diagnosing and examining before surgery of epilepsy. However, EEG which has the good temporal resolution, has the poor spatial resolution. For this reason, magnetic resonance imaging (MRI) is used to diagnose, treat, and determine a structural lesion that can be treated by surgery for epilepsy. This project(CUBAP-TEKNO/017) has been developed with the Departments of Software Engineering, Computer Engineering, Neurology and Radiology in Cumhuriyet University. It is aimed to combine EEG with good temporal resolution and MRI with good spatial resolution into one file format for diagnosis and treatment of epilepsy patients. In this tool, MRIs of the patient are segmented such as the skull, cerebrospinal fluid, white matter, gray matter, and brain with using 2D Wavelet Transform. The epileptic activities of EEG are also displayed on segmented MRI of the same patients. The epileptic EEG signals and their reports are hidden on segmented MRI by a data hiding technique. Experts are able to see EEG activity on the segmented brain while navigating MRI. The performance between stego image and cover image with EEG data is measured by statistical analysis methods such as the mean square of error (MSE), peak signal-to-noise ratio (PSNR), structural similarity measure (SSIM), universal quality index (UQI), and correlation coefficient (R).

CLASSIFICATION AND SEGMENTATION TOOL OF BRAIN MAGNETIC RESONANCE IMAGES WITH DEEP LEARNING

ABSTRACT:

Diagnosis, early detection, and accurate estimation of the stage in disease are important for treatment. Furthermore, accurate estimation of the brain tumors’ borders also plays a key role in treatment planning. Brain tumor segmentation involves the differentiation of tumor tissues from normal brain tissues such as gray matter, white matter, and cerebrospinal fluid. In this project, it is aimed to classify and segment the tumors in T1-weighted brain magnetic resonance (MR) images of the patients with using deep learning to assist the experts in diagnosis and treatment.

DetaILS...