Wavelet Transform and WCA Based Deep Convolutional Network for Brain Tumor Detection and Classification from Magnetic Resonance Images

Abstract

Deep learning is playing vital role in the research of neuroscience for studying brain images. It is the one of the most frequently used diagnosis tool to classify irregularities in the brain. Classifying the type of tumor from Magnetic Resonance (MR) images in the diagnostic system depends on size, shape, and position of tumor which varies from different patient’s brain. Many efforts have been made for image detection and classification, but getting accuracy in classification and detection is a challenging task. Motivated by the complexity of classification brain tumors, this  paper presents a WCA (Sine Cosine Algorithm) based DCNN (Deep Convolutional Neural Network) model has been developed for classification of brain tumors into malignant (cancerous) and benign (noncancerous) category. The discrete wavelet transform segmentation technique has been utilized to improve the performance of detection process. The Daubechies wavelet is considered for segmentation process and shows its improved capability in performance during detection of brain tumor categories from the magnetic resonance images.  Further, the segmented images are given as input to the WCA-DCCN model for classification cancerous and non-cancerous tumors. The WCA clones the sea and rivers in the process of water cycle and through this algorithm the optimization of the weights of the DCCN has been made to improve the performance of the conventional DCCN. Also the different category of hidden neuron functions at the hidden layer has been tested with the new hybrid WCA-DCCN model and comparison results are presented. DCCN+WCA shows an accuracy of  98.93%  and 97.23 % during training and testing.