120

Bioinformatics of the Brain

FIGURE 4.2

Original and skull stripped MRI images. The original MRI image featured in

this figure was selected from the dataset available as open source on Kaggle

[5].

In the literature, different researchers have used various filter types in the

pre-processing phase of brain tumor detection. Dandıl proposed a computer-

aided approach for the detection of brain tumors [9]. In the pre-processing

stage of this approach, the arithmetic median filter was employed to improve

the quality of the MRI image, followed by the Laplacian filter to sharpen the

tumor boundaries, and finally, the histogram equalization filter was applied

to minimize contrast differences. Malathi and Kamal [16] utilized median and

Gaussian filters to remove the noise and sharpen MRI images respectively in

the pre-processing stage of their study where they detected brain tumors using

the k-means clustering technique. Dhage et al. [17] employed bilateral and

median filters to eliminate noise from the pre-processing module of the system

they suggested for the detection of brain tumors. Based on the results of the

performance comparison, the median filter demonstrated better performance.

Reddy et al. [18] and Zotin et al. [19] utilized a median filter to reduce the noise

and filter out unwanted pixels in MRI during brain tumor detection studies.

Aslıyan and Atbakan [20] utilized Gaussian and Wiener filters to improve the

quality of the MRI image in their automatic brain tumor segmentation study.

4.3.2

Skull Stripping

Identifying the area that has to be processed and cleaning the other portions

are prerequisites before working on images. A skull stripping operation is a

procedure used to remove the skull and any surrounding areas from MRI scans

of the brain. To ensure the successful completion of subsequent stages, it is

essential to define the correct and accurate boundaries of the brain area. An

example of a skull-stripping operation is illustrated in Figure 4.2. The figure

displays the original MRI image on the left side, while the right side illustrates

the appearance of the original MRI image after the skull separation procedure.

A study emphasizing the importance of skull stripping in the detection of

brain tumors was carried out by Aslıyan and Atbakan [20]. The success of