×

# Optimization of Various Analytical and Iterative Image Reconstruction Methods

## M Mahdian Manesh*, R Faghihi , Shiraz universityShiraz

• M Mahdian Manesh

# Presentations

(Sunday, 7/14/2019) 4:00 PM - 4:30 PM

Room: Exhibit Hall | Forum 9

Purpose: The purpose of this research is to change the various parameters of image reconstruction algorithms of filter back projection, ART and SIRT to achieve better CNR and SNR

Methods: CT scan images are reconstruct with two analytical or iterative approaches. In an analytical approach, a mathematical model is proposed for the relationship between the unknown values (image matrix) and the values of the sinogram matrix, then information is extracted from the projection data and the image matrix is completed. In the iterative approach, first, the arbitrary value is selected as the pixel attenuation coefficient, then in later stages the values are corrected in such a way that the sum of the coefficients of attenuation corresponds to the amount of attenuation coefficient along each projection. In this research, two analytical and iterative approaches are used. In the analytical approach, the data was reconstructed by the filter back projection algorithm with different filter (Hann, Cosine, Ram-Lak, Shepp-Logan and Hamming) coded in MATLAB software. In the iterative method, ART and SIRT algorithms were implemented in MATLAB software. 8 different Iterative methods, taking into account the effect of various parameters such as relaxation factor, pixel weighing factor, and order of solving equation problems and how to select these parameters, were created. Finally, all of the above algorithms applied on the results of practical data was given by A CT scan. Then created images were compared in terms of SNR, CNR, and reconstruction time and the optimal algorithm was selected for image reconstruction

Results: Among different methods of implementing repetitive algorithms, ART's implementation method provides similar results.for SIRT, ClassicalLandweber is better. In FBP, the Han filter also gives good results

Conclusion: According to the results the FBP (Hann filter) and SIRT (Classical Land weber) methods, have a better SNR and CNR than other methods.

Keywords

Not Applicable / None Entered.

Taxonomy

IM- CT: General (Most aspects)

Contact Email