Brain Mri Images Quality Enhancement Using Image Processing Techniques

Vispalingam, Asvhini Seema (2023) Brain Mri Images Quality Enhancement Using Image Processing Techniques. PhD thesis, Universiti Sains Malaysia.

Download (436kB) | Preview


Image quality of MRI images is important for diagnostic purposes. In this research, MATLAB software version R2016a is used for image pre-processing and processing operations which includes image enhancement, noise reduction, and morphology image processing and thresholding ad edge segmentation techniques. The images were analysed quantitatively and qualitatively in this experiment. ImageJ software was used to analyse the images quantitatively. This experiment was done using two brain tumour MRI images. The images are acquired from an online platform; The Cancer Images Archive with public access. The two patient’s identities were kept discreet in the TCIA domain. Image 1 is of a female patient of 63 years old and image 2 is of a male patient of age 58 years old. The first step in image pre-processing is filtration with median and mean filters. Filtration aims to reduce the noise present in images. Two filter masks were used for mean filtering. They are 3×3 mean filter mask and 4×4 mean filter mask. Three image contrast enhancement techniques were studied in this experiment. They are the manual and MATLAB contrast stretching method, histogram equalization method and Contrast Limited Adaptive Histogram Equalization (CLAHE) method. The thresholding technique was applied to the images for the qualitative analysis of the images. The thresholding was conducted according to Otsu’s thresholding method. In this experiment, several threshold values were experimented with to obtain the best image thresholding.

Item Type: Thesis (PhD)
Subjects: Q Science > QC Physics > QC1 Physics (General)
Divisions: Pusat Pengajian Sains Fizik (School of Physics) > Thesis
Depositing User: Mr Mohammad Harish Sabri
Date Deposited: 13 Mar 2024 03:53
Last Modified: 13 Mar 2024 03:53

Actions (login required)

View Item View Item