Tez özetleri Astronomi ve Uzay Bilimleri Anabilim Dalı

Medical Image Modelling Via Sympes Algorithm
Medical image transmission and archiving are widespread in modern communications systems. Minimising the amount of information required to diagnosability reconstruct a medical image could significantly increase the capacity of digital medical image transmission and archiving systems.
In this thesis, a novel method based on generation of the so-called classified energy and pattern blocks (CEPB) is introduced and evaluation results are presented. The most significant property of this method is to generate any medical image independent from any image obtained from several imaging (MR, CT, CR) modality. Furthermore, the new method provides substantial compression ratio with respect to human visual system.
In the proposed method, two different vector sets called Pre-defined Vector Banks, CEPB, are constructed from any patients medical image obtained by whatever imaging technique. Generally in image compression models due to large image sizes, images are divided into ixj blocks for ease of mathematical operations. In this work medical images are divided into ixj blocks. Every block is modelled as .

Here, P_IP is the pattern vector which defines the edges and details of the block, E_IE is called as energy vector which defines the luminance level of the block, G_i , is the scaling coefficient of the image block.

In the method first CEPB is constructed and then it is located at both the transmitter and receiver sides of the communication system. CEPB contains almost optimum forms of the block vectors which can define E_IE and P_IP in the least mean square sense (LMS). As an encoding process then the energy and pattern blocks of the input images to be reconstructed are determined with the same way in the construction of the CEPB. This process is also associated with a matching procedure to obtain the index numbers of the classified energy and pattern blocks in CEPB which best represents or matches the energy and pattern blocks of the input images. The encoding parameters are the block scaling coefficient (BSC) and the index numbers of energy and pattern blocks (I_E, I_P) determined for each block of the input images.
In the decoding process, all these parameters are sent through the communication channel from the transmitter part to the receiver part and the classified energy and pattern blocks associated with the index numbers are pulled from the CEPB. Then the blocks of the input image are reconstructed in the receiver part using a mathematical model that is proposed.
In this work, evaluation results are divided into two parts. In the first step results are examined according to PSNR values. Afterwards, radiologists diagnose images with maximum and minimum PSNR values.
In the second part of this study, input images are reconstructed from 5 different database (CEPB) and combinations of them, PSNR values are examined and highest and lowest PSNR valued images are diagnosed by radiologists.
In the last part of this thesis, block-scaling coefficients, classified energy and pattern block’s effects are inspected and diagnosed by radiologists.