Enhanced Early Brain Tumor Detection Crossing Blood-Brain Barrier through MRI Images Using Berkeley Wavelet-Transformation-Based Segmentation

Abstract

Brain tumor is one of the reasons for several mortality cases in hospitals. Early detection and diagnosis of brain tumors are necessary to cure the disease early. The extraction of the tumor from the brain's magnetic resonance image (MRI) is considered to be a difficult task when done by clinical experts, and it is also pretty time-consuming. These drawbacks can be overcome by using computer vision-based technologies. The proposed method detects brain tumor crossing the blood-brain barrier (BBB) through MRI images by using Berkeley wavelet transformation (BWT) for segmenting the affected areas. Support vector machine (SVM) is used for classification purpose by which the classification process is divided into two different categories namely, the tumor affected and tumor non-affected parts. Initially, the acquired image is converted to a greyscale from RGB. Then, image segmentation is done. During the image segmentation, morphological operations are carried out. Two morphological operations have been used in the proposed system. They are erosion and dilation. Both these techniques are used for edge detection. In erosion, the pixels are removed from the edges of the tumor image. In dilation, pixels are added at the edges of the tumor images. After the morphological operation, feature extraction is carried out. The features like homogeneity, contrast of the image and the energy might be determined. Then, the image is classified using the SVM classification algorithm. The experimental results have been tabulated and depicted using graphical representations. Comparing to the existing approaches the proposed method is proved to be better in accuracy and efficiency.