A study of classification and feature extraction techniques for brain tumor detection
- PDF / 2,304,506 Bytes
- 20 Pages / 595.276 x 790.866 pts Page_size
- 31 Downloads / 193 Views
TRENDS AND SURVEYS
A study of classification and feature extraction techniques for brain tumor detection Vatika Jalali1
· Dapinder Kaur1
Received: 28 July 2020 / Revised: 10 October 2020 / Accepted: 16 October 2020 / Published online: 12 November 2020 © Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Medical imaging aids in the analysis of interior parts of the human body such as the functioning of the organs or tissues for early treatment of diseases. Many different types of medical imaging technologies exist, for example, X-ray radiography, magnetic resonance imaging, endoscopy, positron emission tomography, CT scan (computed tomography), and many more. A tumor is an abnormal tissue in the brain which causes damage to the functioning of the cell. Therefore, brain tumor detection is an incredibly tricky task. Manual detection of a tumor is quite risky as it involves the insertion of a needle in the brain. Thus, there is a need for automated brain tumor detection systems. The well-timed detection of the tumor can add to accurate treatment and can increase the survival rate of patients. From machine learning techniques, namely K-nearest neighbor, support vector machine, and more to soft computing techniques, namely artificial neural network, self-organizing map, and others hold a significant stand in detection and categorization of brain tumor. Various methods including deep learning-based classifiers such as convolutional neural network, recurrent neural network, deep belief network (DBN), and others are used to make it easier to detect the tumor. Hybrid classifiers were also used for classification systems such as combining the machine learning approach with soft computing. This study is to summarize and compare the work of various authors on automatic brain tumor detection using medical imaging. Based on the accuracy, specificity, and sensitivity parameters, the results of different techniques are analyzed and compared graphically. Keywords Tumor · MRI · Feature extraction · Classifier
1 Introduction Digital image processing is one of the emerging techniques that have been explicitly used for past decades. It is a method to process a digital image through an algorithm to enhance the image to extract useful information. A digital image is a two-dimensional representation of a real image. It consists of a finite number of pixels. Various noise removing filters and contrast enhancement algorithms are used to improve the quality of the image. Enhanced image delivers the skillful information for further processing [1]. The input to the computer-aided system is the digital image which is further processed by using a large number of algorithms and pro-
B
Vatika Jalali [email protected] Dapinder Kaur [email protected]
1
Computer Science, Chandigarh Engineering College, Landran, Punjab, India
duces the required output. A range of algorithms is used for performing various operations such as image sharpening, zooming, blurring, noise removal, edge detection, and many more [2]. The
Data Loading...
