Restoration Analysis of Various Type of MRI Brain Tumor Using Blind De-Convolution Pre-Processing Technique

摘要

In radiology department, original patient images may have poor quality, blurriness and noisiness, due to this problem doctor can suffer from the initial processing of diagnosing. We proposed blind deconvolution based PSF estimation using EM technique with an estimated parameter for reducing blurry, noisy in MRI brain tumor degraded images. The brain tumor MRI patient images are used for the experimental work. These images include-(Astrocytoma, Ganglioglioma, Glioblastoma, Epidermoid, Mixed Glioma and Meningnet) (5type of tumors) were used for the experiment. Brain tumor image restored using BD methodology based PSF estimation using EM methodology. The experimental result shows that proposed methodology was found to be 0.98 sec with 10 iterations and regularized lucy method [5] found to be 1.524 sec with 103 iteration to complete the restoration process. The image quality measurement value is better than regularized lucy method. The proposed methodology measures the image quality assessment PSNR value is higher with higher value and higher SSIM.