Using Singular Value Decomposition in a Convolutional Neural Network to Improve Brain Tumor Segmentation Accuracy

A brain tumor consists of cells showing abnormal growth. The area the significantly affects choosing type treatment and following course disease during treatment. At same time, pictures Brain MRIs are accompanied by noise. Eliminating existing noises can impact better segmentation diagnosis tumors. In this work, we have tried using analysis eigenvalues. We used MSVD algorithm, reducing image noise then deep neural network to segment in images. proposed method's accuracy was increased 2.4% compared original With Using method, convergence speed has also increased, effectiveness.