Research Area:  Machine Learning

Abstract:

Many studies have been conducted to examine abnormal conditions in brain structures and to detect Alzheimer-s and Dementia states using features derived from medical images. From these data, it is very important to detect the diagnosis of Alzheimer-s and Dementia disease early and to provide appropriate treatment to the patients. Quality magnetic resonance (MR) images are requested to make this diagnosis. But while producing a quality image, it also brings less spatial coverage and longer scanning and identification time. In this context, biomedical image processing has undergone a serious expansion and has become an interdisciplinary research field that includes many fields. Computer Aided systems have become an important part in the diagnosis process. With the development of computer aided systems, producing quality information for the diagnosis of disease in image processing applications has caused various problems. Such difficulties are tried to be overcome with artificial intelligence technology and super-resolution (SR), which has gained great importance in image processing lately. Using the super resolution methodology, a high resolution image is obtained from the low resolution image. Thus, the image processing timing is shortened and an image with desired features can be obtained. This shortens the irritating and long-lasting MR imaging process. In addition, it provides convenience for the diagnosis of the disease with the improvements it provides on MR images. Recovering the image is an important step in this process. The quality of the reconstructed image depends on the restoration methods. The functionality of artificial intelligence technology in image processing and biomedical fields is increasing day by day. The deep learning method is preferred in techniques aimed at obtaining a reconstructed quality image. At the same time, various artificial intelligence methods are widely used for classifying and detecting the data obtained. One of the most common of these is neural network (NN) methods.. Deep learning, a special method of neural networks, is widely used in classification methods due to its superior structural properties. When studies are examined, it is seen that DL methods are widely used. The success of the proposed methods is increasing day by day.

Keywords:  

Author(s) Name:  Emre Altinkaya, Kemal Polat, Burhan Barakli

Publisher name:  IEC Science

DOI:  10.33969/JIEC.2019.11005.

Volume Information:  Volume 1, pages: 39-53

Paper Link:   https://iecscience.org/jpapers/55