Brain Tumor Detection System Based on Sending Email Using Gray Level Co-Occurrence Matrix and Back-Propagation Neural Network

Abstract

The human brain is the main organ in regulating and coordinating most movement’s behavior and body functions. The brain is a vital organ of human life because it has an important function as the control center for thousands of body activities. Related to diseases that can attack the human brain is a tumor. A brain tumor is abnormal of proliferation cells in brain tissue that can grow uncontrollably. Early detection of brain tumors is important to check the severity degree of the tumor. Lateness for detecting brain tumors will be fatal because of the death risk. Early detection of brain tumors can decrease the severity degree because the patient can immediately get treatment. The proposed method for detecting the brain tumor uses the Gray Level Co-occurrence Matrix feature, which will produce a Contrast, Homogeneity, Energy, and Correlation value and the classified method using the Backpropagation Neural Network algorithm. Carried out the detection of tumors minicomputer use Raspberry PI 4B, which will send a message and information via email. The results of highest accuracy with Gray Level Co-occurrence Matrix parameter of distances (d) = 1,2,3,4, and angles (θ) =0o, 45o, 90o and 135o. The highest accuracy parameter d distance = 3, and angle (θ) = 90 is 0.701 and Classification of Neural Network Backpropagation (BPNN) uses layers 1, 3 and 5 for the highest accuracy is the hidden layers = 5, fixed in the architecture of Back-propagation Neural Network (BPNN) input layers= 4,hidden layers= 5 and output layers= 2 with an accuracy of 0.871. Then the result is sent by email. Computation Time for tested dataset image has average = 0.601. Performance results in deep learning are accuracy = 88.3, precision = 82.8 and sensitivity = 92.3.