Caries detection in the dental panoramic X-ray images using region based segmentation and support vector machine

Other Title(s)

اكتشاف تسوس الأسنان في صور الأشعة السينية البانورامية باستخدام التجزئة المعتمدة على المنطقة و آلة متجه الدعم

Dissertant

Muhammad, Muna Khalid Jasim

Thesis advisor

al-Kabnah, Khalid Abd al-Hafiz

University

Amman Arab University

Faculty

Collage of Computer Sciences and Informatics

Department

Department of Computer Science

University Country

Jordan

Degree

Master

Degree Date

2019

English Abstract

With the rapid development of digital image processing, many approaches were proposed to automatically detect diseases in the human body through processing medical images.

There are two kinds of medical images are used to diagnose mouth diseases, extra-oral radiographic (dental panoramic X-ray) image and intra-oral radiography (Bitewing and Preipical) images.

These images are used by dentists to diagnose different mouth diseases, that are related to gums, teeth and the bones of the upper jaw and lower jaw.

This research will focus on detecting dental caries in dental panoramic X-ray images, which is one of the most popular of mouth diseases.

It is very challenging to use a computer to diagnose dental panoramic X-ray images, that is because of the complex nature of these images.

Dental panoramic X-ray images include a lot of details about the area of the gums teeth from other parts because the bones and teeth have the same color intensity in the images.

This research proposes and implements an approach to detect dental caries in dental panoramic X-ray images.

The proposed approach uses two well-known handed crafted features, Scale-Invariant Feature Transform (SIFT) and Histogram Oriented Gradient (HOG) to extract the teeth features.

Then, the extracted features are encoded using the Fisher vector.

Finally, Support Vector Machine (SVM) is used to classify the extracted features into two classes decayed and healthy teeth.

In addition, another experiment was conducted by fusion the two features SIFT and HOG, the results showed that the fusion process can enhance the classification accuracy.

The experimental results show that the accuracies of using HOG, SIFT and the fusion of HOG with SIFT for dental caries detection are 80.3%, 89.7% and 90.3%, respectively.

An experiment was performed to compare the performance of four well-known edge detection techniques: Sobel, Canny, Roberts and Prewitt.

Mean Square Error (MSE), Mean Absolute Error (MAE) and Peak Signal-to-Noise Ratio (PSNR) metrics were used to make comparisons between these edge detection approaches.

Another experiment is in order to evaluate the performance of the Region-based segmentation approach to segment different parts in dental panoramic X-ray image.

Main Topic

Information Technology and Computer Science

No. of Pages

77

Table of Contents

Table of contents.

Abstract.

Abstract in Arabic.

Chapter One : Introduction.

Chapter Two : Literature review.

Chapter Three : The procedures and methodology.

Chapter Four : Results and analysis.

Chapter Five : Conclusions and future work.

References.

American Psychological Association (APA)

Muhammad, Muna Khalid Jasim. (2019). Caries detection in the dental panoramic X-ray images using region based segmentation and support vector machine. (Master's theses Theses and Dissertations Master). Amman Arab University, Jordan
https://search.emarefa.net/detail/BIM-932645

Modern Language Association (MLA)

Muhammad, Muna Khalid Jasim. Caries detection in the dental panoramic X-ray images using region based segmentation and support vector machine. (Master's theses Theses and Dissertations Master). Amman Arab University. (2019).
https://search.emarefa.net/detail/BIM-932645

American Medical Association (AMA)

Muhammad, Muna Khalid Jasim. (2019). Caries detection in the dental panoramic X-ray images using region based segmentation and support vector machine. (Master's theses Theses and Dissertations Master). Amman Arab University, Jordan
https://search.emarefa.net/detail/BIM-932645

Language

English

Data Type

Arab Theses

Record ID

BIM-932645