Deep Learning Techniques for Prediction, Detection, and Segmentation of Brain Tumors

Deep Learning Techniques for Prediction, Detection, and Segmentation of Brain Tumors

Prisilla Jayanthi, Muralikrishna Iyyanki
DOI: 10.4018/978-1-7998-3591-2.ch009
OnDemand:
(Individual Chapters)
Available
$33.75
List Price: $37.50
10% Discount:-$3.75
TOTAL SAVINGS: $3.75

Abstract

In deep learning, the main techniques of neural networks, namely artificial neural network, convolutional neural network, recurrent neural network, and deep neural networks, are found to be very effective for medical data analyses. In this chapter, application of the techniques, viz., ANN, CNN, DNN, for detection of tumors in numerical and image data of brain tumor is presented. First, the case of ANN application is discussed for the prediction of the brain tumor for which the disease symptoms data in numerical form is the input. ANN modelling was implemented for classification of human ethnicity. Next the detection of the tumors from images is discussed for which CNN and DNN techniques are implemented. Other techniques discussed in this study are HSV color space, watershed segmentation and morphological operation, fuzzy entropy level set, which are used for segmenting tumor in brain tumor images. The FCN-8 and FCN-16 models are used to produce a semantic segmentation on the various images. In general terms, the techniques of deep learning detected the tumors by training image dataset.
Chapter Preview
Top

Deep Learning Techniques

This approach is used to build and train neural networks (NN), with high propitious decision-making nodes. An algorithm is said to be deep if the input data is passed through a series of layers before it gives output. The state-of-the-art of DL is focused on training NN models using the backpropagation algorithm. Deep Neural Networks (DNN) is one among the DL technique and all the networks that have more than a two or three layers are DNN. The most popular DL techniques are shown schematically in figure 1 (Brownlee, 2019):

DL has been dominant in the medical imaging, in which different techniques are used for various analysis purposes. One of such technique of DL named ANN works on the numerical datasets of the patient in healthcare whereas CNN deals with the medical imaging that includes magnetic resonance imaging, digital images and so on.

Figure 1.

Techniques of Deep Learning

978-1-7998-3591-2.ch009.f01
  • The four tasks of DL are:

Image Classification, that includes two types of classification namely binary and multiclass. It allocates a label to an image.

  • Labeling an x-ray comes under binary classification whereas

  • Naming a face in the photo is multiclass.

Image Classification with Localization (Object Localization) includes class labelling and displaying the object location by a bounding box.

  • Classifying different objects like birds and drawing a rectangular box known as bounding box around the birds.

  • Labeling a digital image in MRI as cancer with a box around the region.

Object detection is the task of image classification with localization. Whenever an image has multiple objects and different objects are to be detected; it requires localization and classification.

  • Labeling each element in the satellite image with a bounding box.

  • A bounding box and labeling each vehicles, and pedestrians on a road.

Image segmentation is the technique of drawing a line around the object detected in the image. The process of dividing an image into segments is called as Image Segmentation. The segmentation identifies the specific pixels in the image that belong to the object image with a fine-grained localization (Brownlee, 2019).

Key Terms in this Chapter

Object Detection: It deals with detection of semantic objects of a certain class (such as humans, buildings, or cars) in digital images and videos.

Neural Network: It is a circuit of neurons and with layers of a modern sense.

Algorithm: A step-by-step approach for solving a problem.

Greyscale: The image has the value of each pixel with the amount of light that carries only intensity information.

Perceptron: A perceptron is a single-layer neural network. It includes input values, weights and bias, net sum, and an activation function.

Bias: An additional parameter used to tune the output along with the weighted sum of the inputs to the neuron.

Neurons: It is a cell and specialized to transmit information to the entire system.

Weights: They are numerical parameters in an ANN that converts an input to control the output.

Image Classification: It groups items based on the categories.

Image Segmentation: It is the process of dividing an image into multiple segments. The aim is to get more meaningful information and easier for analyzing.

Thresholding: The method is based on a threshold value to turn a gray-scale image into a binary image.

Image Processing: It uses the computer algorithms to perform image processing on digital images.

Complete Chapter List

Search this Book:
Reset