Brain Tumor Detection System

Project Title

Brain Tumor Detection System

Project Area of Specialization Artificial IntelligenceProject Summary

Computer-aided diagnosis (CAD) systems can enhance the diagnostic capabilities of radiologists and physicians, optimize the accuracy .Also it can reduce the time required for accurate classification of brain tumor.  Brain Tumor is one of the major causes of death in recent years and becoming life-threatening. There are multiple types of brain tumor types exist and each tumor type has different structure ,placement in head and different nature.Correct Diagnosis of brain tumor type can effect the planning of treatment and increase the survival rate. Deep learning and classification through neural network  playing the major role in computer vision field . Neural network work better in learning the structure of tumor and predicting the tumor class precisely .The reliable and automatic classification scheme is essential to prevent the death rate of humans.The automatic classification of brain tumors is a very challenging task in the great spatial and structural variability of the surrounding region of the brain tumor. In this CAD project, we propose the automatic detection of brain tumors by using the classification of convolutional neuronal networks (CNN) .We are intended to develop a CAD that enhance and optimize the tumor diagnostic process. It will help radiologists for decision making and further surgical treatments. 

Project Objectives

We aim to develop a CAD to diagnose and detect the tumor efficiently in brain MRI image.  The main objective of this project is to build a system which will helpful for the radiologists and physicians to classify the brain tumor accurately and efficiently in MRI image.We also intend to accomplish following sub-objectives: use in real time applications , fast computational time , less prone to human error, robust and flexible to large data amount ,compare accuracy results with other techniques ,cost effective .

Project Implementation Method

In this project we used machine learning techniques and artificial intelligence. Data is gathered from patients of Brain Tumor, their MRI images are collected and formed a huge dataset.It consists of 3064 T1 CE-MRI images. The images of 233 patients  are collected from different sides of brain: backside, front side and top . There are 994 axial images , 1045 coronal images and 1025 sagittal images captured.This dataset is given as input to system. First of all pre-processing occurs which means your data should be clean , well specified  and  consistent. All the images are resized and all images having extension .jpg. Whole dataset is split into two parts, 80%testing and 20% training. Then classifier CNN will be used to extract features and classification of brain tumor. At the end result is generated. The dataset i used is of Jun Cheg .

Benefits of the Project

It is extended to heigher level to provide many benefits such as:

•It can enhance the diagnostic capabilities of radiologists and physicians, optimize the accuracy.

•It can reduce the time required for accurate classification of brain tumor.

•Correct Diagnosis of brain tumor type can effect the planning of treatment and increase the survival rate.

•This system can reduce the work-load and  also avoid the mistakes that can be done by the radiologist.

Technical Details of Final Deliverable

Technical details include understanding basics of neural networks, their implementstion in python, understanding working of libraries Keras, tensorflow, openCv. Understanding Deep learning techniques and Convolutional Neural Networks etc.

Final Deliverable of the Project Software SystemCore Industry MedicalOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development Goals Good Health and Well-Being for PeopleRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	23000
GPU	Equipment	2	8000	16000
RAM	Equipment	2	2000	4000
Documentation Printing	Miscellaneous	300	10	3000