Brain Tumor Classification Using Convolutional Neural Network

Brain tumor is a kind of cancer, in which tissues in the brain grows rapidly and unevenly in the brains and causes huge threats on human life. Brain tumor is recognized as one of the common dreadful cancers among adults and it also affects the children too. This kind of cancer is categorized into two types, such as benign tumor and malignant tumor. However, benign tumor is curable, whereas recovering of patients whoever affected by malignant tumor has less chance to survive. Nowadays, MR images are usually employed to detect the kinds of brain tumor. Early classification and identification of tumor is significant to treat the tumor and saves the human life from early death. However, the classification of brain tumor and percentage in change detection using pre-operative and post-operative MR images is a very challenging task. In order to overcome such issues, this research proposes a new effective technique for brain tumor classification and determination of pixel change detection using proposed Deep Belief Network (DBN) + Deep Convolutional Neural Network (DCNN). The process involves four phases, such as pre-processing, segmentation, feature extraction, and classification. The combination of DBN + CNN is employed for decision making based on error function. The DBN + CNN are trained utilizing the developed BirCat algorithm. Moreover, the proposed approach achieved a maximum accuracy of 0.957, sensitivity of 0.967, and specificity of 0.918.

Download Full-text

Brain tumor classification in MRI image using convolutional neural network

Mathematical Biosciences and Engineering ◽

10.3934/mbe.2020328 ◽

2020 ◽

Vol 17 (5) ◽

pp. 6203-6216

Author(s):

Hassan Ali Khan ◽

◽

Wu Jue ◽

Muhammad Mushtaq ◽

Muhammad Umer Mushtaq ◽

...

Keyword(s):

Neural Network ◽

Brain Tumor ◽

Convolutional Neural Network ◽

Tumor Classification ◽

Mri Image

Download Full-text

Convolutional Neural Network Based Brain Tumor Classification Using Robust Background Saliency Detection

Journal of Medical Imaging and Health Informatics ◽

10.1166/jmihi.2021.3849 ◽

2021 ◽

Vol 11 (10) ◽

pp. 2610-2617

Author(s):

K. Uthra Devi ◽

R. Gomathi

Keyword(s):

Neural Network ◽

Brain Tumor ◽

Convolutional Neural Network ◽

Saliency Detection ◽

Automated Analysis ◽

Regions Of Interest ◽

Tumor Classification ◽

Tumor Section ◽

Canny Edge ◽

The Brain

To perceive the tumors found in brain and their treatment, experts manually note and identify different Regions of Interest (ROI). To overcome the faults and divergences during this state, automated analysis is performed. A unique technique is used to classify the tumor section of the brain from an MRI is proposed using saliency-focused image depiction and optimization in classification based on CNN. Primarily, the MRI images are pre-processed using the Canny Edge Finding algorithm and then those images are represented as saliency driven based on Robust Background Saliency Detection (RBD). Followed by the abstraction of features then classifying the image is performed using CNN along with ADAM optimization. The implementation is accomplished, and the results are analyzed, showing that it outperforms previous techniques.

Download Full-text