Cognitive Analytics and Reinforcement Learning

12Investigation of Feature Fusioned Dictionary Learning Model for Accurate Brain Tumor Classification

P. Saravanan¹, V. Indragandhi², R. Elakkiya³ and V. Subramaniyaswamy⁴*

¹School of Computer Science and Engineering, Vellore Institute of Technology, Chennai, India

²School of Electrical Engineering, Vellore Institute of Technology, Vellore, India

³Department of Computer Science, BITS Pilani, Dubai Campus, Dubai, United Arab Emirates

⁴School of Computing, SASTRA Deemed University, Thanjavur, India

Abstract

Brain tumors, which are one of the most prevalent and deadliest diseases, have an extremely low life expectancy when they are at their most advanced stage. The fate of brain tumors may be significantly influenced by early diagnosis. Brain tumors are diagnosed from MRI images using standard symptoms. However, using an MRI to determine the type of tumor is time-consuming, challenging, and error-prone as the number of cases increase. Given the significant anatomical and geographical heterogeneity of the brain tumor’s surrounding area, automatically classifying brain tumors is a highly difficult process. In this work, we proposed a deep learning approach that considers the hybridization of CNN and LSTM with Dictionary Learning. Further, the proposed model is optimized using Feature Fusion. The experimental results prove that the proposed Feature Fusioned Dictionary Learning Model outperforms in terms of accuracy and improves the classification process by 4.58% from the core model. ...