International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 06 Issue: 04 | Apr 2019 www.irjet.net p-ISSN: 2395-0072 © 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 4323 An Efficient Brain Tumor Detection system using Automatic segmentation with Convolution Neural Network Sadaf Naz 1 & Nitesh Kumar 2 1 Research Scholar, Dept. of Electronics & Communication, Sagar Institute of Research Technology & Science, Bhopal, Madhya Pradesh 462041, India 2 Assistant Professor, Dept. of Electronics & Communication, Sagar Institute of Research Technology & Science, Bhopal, Madhya Pradesh 462041, India ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract - Brain is one of the most complex organs in the human body that works with billions of cells. A brain tumor is a collection, or mass, of abnormal cells in your brain. This can cause brain damage, and it can be life- threatening, so an early detection of tumor is required and for that a reliable technique is required, thus With the advancement in the technology the concept of image processing play an important role in medical field and with convolution neural network we get better promising results. In this paper we use the methodology in which first we filter the image by median filter then automatic segmentation by Otsu then morphological operation for filtration and dilation then classification done by CNN. We prepared the brain MRI dataset and performed the methodology using MATLAB R2015a Weka 3.9 tool was used for performing the classifications .The evaluation of the performance for the proposed methodology was measured in terms of average classification rate, average recall, average precision, PSNR and MSE with accuracy. Key Words: (Size 10 & Bold) Key word1, Key word2, Key word3, etc (Minimum 5 to 8 key words)… 1. INTRODUCTION In this research paper with the advancement in the technology the concept of image processing play an important role and especially with the concept of artificial neural network. And for efficient detection of brain tumor that is a medical field ,we use the image processing concepts with deep learning that give better promising results in the different field, for example, speech recognition, handwritten character recognition, image classification, image detection and segmentation and disease detection. Brain Tumor Symptoms: Symptoms (signs) of benign brain tumors often are not specific. The following is a list of symptoms that, alone or combined, can be caused by benign brain tumors; unfortunately, these symptoms can occur in many other diseases: vision problems ,hearing problems ,balance problems, changes in mental ability (for example, concentration, memory, speech), seizures, muscle jerking, change in sense of smell ,nausea/vomiting, facial paralysis, headaches, numbness in extremities. The research is based on the automatic detection of brain tumor and the classification is done by using deep convolution neural network. In this paper, the main objective is to develop a high speed classifier using CNN to classify the abnormality of brain. We propose architecture for automatic brain tumor detection with preprocessing filtering object separation and segmentation with morphological operations. Our network is based on Mina Reza et al. [2] with some improvements and enhancements. First we take the input of brain MRI images. Then preprocessing, filtering is required for filtration. The Otsu segmentation system is use for location of irregular tissues inside the tumor. In the segmentation output, the size and shape of the tumor is shown. The textural and intensity base features are drawn. The CNN using stack auto encoders is used as a classifier. We stack and the softmax layers are used to construct the deep network with 10 hidden layers. Surgery is the usual first treatment for most brain tumors. Before surgery begins, you may be given general anesthesia, and your scalp is shaved. You probably won't need your entire head shaved. Surgery to open the skull is called a craniotomy. The surgeon makes an incision in your scalp and uses a special type of saw to remove a piece of bone from the skull. You may be awake when the surgeon removes part or the entire brain tumor. The surgeon removes as much tumor as possible. You may be asked to move a leg, count, say the alphabet, or tell a story. Your ability to follow these commands helps the surgeon protect important parts of the brain. After the tumor is removed, the surgeon covers the opening in the skull with the piece of bone or with a piece of metal or fabric. The surgeon then closes the incision in the scalp. But apart from that the early detection of tumors is required to save the life and for that we design an automated early detection system 2. LITERATURE REVIEW (1). Umit IIhan & Ahmet IIhan “Brain tumor segmentation based on a new threshold approach” [5].  Median filter.  Threshold based segmentation.  Tumor detection with limited number of parameters.