Abstract:
Computer vision and machine learning are growing elds especially, in medical image
processing. It helps in the research of brain tumor prediction, biopsy guidance, prognosis
monitoring, disease stage identi cation, therapy planning, and therapy response. The
Tumor is the accumulation of abnormal cells. It is the second most common cancer in
children and young people. In the case of Glioma, It is the most common form of the
malignant tumor. which are heterogeneous in nature. starting the diagnosis earlier will
help them to extend their valuable life. Nowadays fully automatic methods have been
able to achieve state-of-art results using Magnetic Resonance Image which can give better
tissue images.
In this thesis, a hybrid algorithm used to detect, classify and segment the brain tumor.
Three main procedures are done in this research they are pre-processing classi cation and
post-processing. Gray level co-occurrence matrix and Discrete wavelet transform used to
excerpt the highlevel ideal attribute from the input. After the classi cation by Hybrid
methodology post-processing is done using a morphological process.
