A Novel Efficient Deep Feature Extractor and Classifier Approach for Brain Tumor Segmentation in Magnetic Resonant Images

The main purpose of this paper is to develop a novel deep feature retrieval approach for performing brain tumor identifying process from Magnetic Resonant images. Brain tumor is caused due to uncontrolled cell divisions. Tumor detection in the early phase is very important which is useful for diagnosis and treatment of tumor. First the input Magnetic Resonant brain image is denoised by using the Modified Decision Based Unsymmetric Trimmed Median Filter (MDBUTMF) and then the image contrast is improved with the Contrast Limited Adaptive Histogram Equalization (CLAHE). After pre-processed the input image the next step is to retrieve the features from the denoised and contrast enhanced image. To extract the features from the pre-processed image this project proposed one novel feature retrieval technique named Deep Efficient Reduced Local Derivative Pattern (DERLDP). After extracting the deep features, the next step is to partition the brain tumor based on these extracted features. To do this process the supervised segmentation approach is employed. Among several supervised segmentation approaches this work uses deep machine learning approach named Convolution Neural Network (CNN). Finally, the extracted features are given as input to these machine learning approach to partition the brain tumor regions. To find out the performance of the proposed deep feature retrieval and deep machine learning approach, four performance metrics are employed namely, Dice Similarity Coefficient (DSC), Positive Predictive Value (PPV), Jaccard Index (JI) and Sensitivity (SEN). From the experimental results, it is shown that the novel DERLDP and CNN performs better than other existing approaches.