Survey on CAD Systems Improving MRI-Based Brain Tumor Classification with Advanced Segmentation and Feature Extraction Techniques

Tumors are among the main causes of death not only in developed countries but also in developing ones. Cancerous neoplasms are a significant health problem in the world. Treatment, thus, calls for early diagnosis. Nearly one in six deaths is caused by cancer, which was approximately 10 million deaths in 2020 given by WHO. Actually, many cancers can be cured if treated effectively and early. In that regard, it presents the CAD system based on designing an automatic detector for detection and classification of brain tumors from MRI images. That is why methodologies, such as Support Vector Machine(SVM), Principal Component Analysis(PCA), Discrete Wavelet Transform and K-means clustering have been used in this paper. It predicts at two levels; first, differentiates between MRI images if they belong to either normal or abnormal types, and then differentiates the cancerous from noncancerous types. Apart from its promising results, there is not enough efficiency in the processing time, though. This paper looks towards noise reduction techniques and image segmentation and feature extraction methods for improving classification performance with the overcoming of limitations of current methods. Further on, advanced computational algorithms, such as the OFCM technique, are discussed to pinpoint more accurate identification of tumor regions. All the image processing and segmentation techniques above with identifying some open challenges have outlined the possible improvement for brain tumor detection systems.