Final Project Proposal : DESIGN OF LUNG CANCER DETECTION SYSTEMS USING X-RAY IMAGE SEGMENTATION AND ADAPTIVE NEURO FUZZY INFERENCE SYSTEMS

My department (Engineering Physics) has five areas of interest, there are instrumentation engineering and control, energy and environmental conditioning, acoustic and building physics, material engineering, and photonic engineering.   This is my final year as as bachelor students, so i have to choose one area of interest for my final project topic. I chose the photonic.  Science of photonic including signal processing, emission, transmission, emission, amplification, and sensing of light. Therefore photonic is very close with both classical and modern optic. My final project proposal's title is DESIGN OF LUNG CANCER DETECTION SYSTEMS USING X-RAY IMAGE SEGMENTATION AND ADAPTIVE NEURO FUZZY INFERENCE SYSTEMS. I can't guarantee that this is the fixed title because i not conduct a seminar proposal yet. But i think it's better for me if i share my project proposal first. I hope this can be useful for readers :D :

1.      Title

DESIGN OF LUNG CANCER DETECTION SYSTEMS USING X-RAY IMAGE SEGMENTATION AND ADAPTIVE NEURO FUZZY INFERENCE SYSTEMS

2.      Summary

Lung cancer is a disease characterised by uncontrolled cell growth in tissues of the lung.. The Process of lung cancer diagnosis depends on several factors : i.e. medical history (personal smoking and secondary expssure, past problem lungs record, current symptoms, activity background, and family history) and physical examination (fever, strange breath sounds, swollen lymph nodes, liver enlargements, hand/ feet/ face/ anklesswelled, changing of skin pigmen, muscle weakness). The results of each step influences the next step in the process.  Imaging test are performed to determine if a cancer cell is present. A chest x-ray is often the first imaging study performed when primary or metastatic lung cancer is suspected. During diagnosis process, subjectivity of the doctor is one of important osbtacles. It is noteworthy that the decision of the doctor is related to the previous diagnosis. Then, to gain the precise diagnosis and interpret the x-ray scan accurately, previous input and output data diagnosis should be automated and used effectively

In this research, a software for detect lung cancer automatically is proposed. This software consict of two main parts : i.e. (1) Image processing software and (2) ANFIS software. Image processing software convert X-Ray scan of the lungs in to digital data, resize and enhance it in to grayscale image format. Lung cancer nodules that present in image are isolated to find their region of interest using image segmentation. Furthermore, In order to convert image as data input in ANFIS software, segmented image  is normalized. In the other hand, ANFIS software is designed to support patients and doctors for lung cancer identification based on imaging test, in accordance with the order of examination. ANFIS software process the normalized image and determined negative or positive value of lung cancer. Furthermore, This results are validated by compare it with doctor diagnosis as human expert or benchmark

3.      Introduction

3.1  Background

The lung is vital organs.  Together with the heart and circulatory system, lung provide sustaining oxygen and rid the body of carbon dioxide.  Normal lungs have a great  capacity to supply the body’s need for oxygen in various circumstancess. Furthermore, lung disease may cause serious problems in respiratory systems and mortality. The Most attacking disease in the lung is lung cancer. Lung cancer is caused by uncontrolled cell growth on these organ [1]. Almost all lung cancer patient are smokers. Tobacco smoke damaged lung cell, causing abnormal cell growth. Although some people who had never smoke get lung cancer, smoking is the leading cause with 90% precentage [2].

Comparing with the other type of cancer, lung cancer is the most causing death for both men and women based on World Health Organization (WHO) data with 19,7% precentage from all cancer [3]. Every year, more than 1.2 million lung cancer case have been diagnosed. People who inhale cigarrete smoke from other smokers (also called as secondhand smokers or passive smokers) also increase lung cancer risk, although they are not smokers.

The lung cancer can be cured easily in initial stage but may be impossible in the advanced stage. In the other hand early detection of lung cancer patient is difficult because it’s prognosis would appear when it comes to advanced stadium. Many of early lung cancers were diagnosed incidentally, after the doctor found symstomps as a result of test performed for an unrelated medical condition. The CT-Scan and X-Ray scan method are used for lung cancer diagnosis in imaging test phase. The CT-Scan displayed more detailed image than X-Ray scan and is able to scan lung organ from many angles. However, more patients chose X-Ray scan because of financial reason [4].

The subjectivity of the doctor is an important obstacles in diagnosing a new patient. It should be noted that the decision of a doctor is related to the last diagnostic. Therefore, to enhance the diagnosis and to interpret the X-Ray scan amore ccurately, the large amount of the empirical input- output data must be automated and used effectively. The diagnosis is similiar with matching procedure that objective is to match each set of the symptoms  (featured space) to a specific case [5]. 

The Fuzzy Neuro system uses procedure of learning to find membership functions which can be expressed from if-then rules. There are many advantages of Fuzzy-Neuro systems : i.e. (1) Allowing incoprporate user experience and the previous knowledge in to classifier, (2) Providing understanding about datasets characteristic, (3) can help for finding the datasets’ independencies, (4) providing explanation which allow users to test the internal logic. Moreover, Adaptive Neuro Fuzzy Inference Systems (ANFIS) can be implemented in many cases such as approximate data, dynamic systems data processing, forecasting goods demands, identifying DNA splice site and image compressions. Various data, e.g. interactions, responses, biomechanical, physical, psychophysical, and psychological parameters are very suitable to be modeled by ANFIS based on the fact that their parameters has highly complex non linear and adaptable systems [5]. A computerised analysis of lung X-ray images can reveal these diseases in their early stages. Most cancer cases start with the appearance of small nodules. This proposal propose the design and implementation of an X-ray image processing to detect early signs of lung cancer using Adaptive Neuro Fuzzy Inference Systems (ANFIS)

3.2  Problem Statements

Based on the background above, the problems in this research can is formulated as follows :

1.      How to process X-Ray scanned image of the lungs for cancer detection purpose using digital image processing?

2.      How to design Adaptive Neuro Fuzzy Inference Systems (ANFIS) for lung cancer early detection in imaging test?

3.3  Scope and Limitation

The main aim of this research is to design image processing and ANFIS systems in imaging test for lung cancer diagnosis. In order to achieve the main aim and the specific objectives of this research, the scope and limitation of the work are described below :

1.      The processed image is collected from lung X-Ray scan

2.      The algorithm designs in digital image processing and ANFIS are constructed using MATLAB R2008a software

3.      The designed system will be used for lung cancer prediction in imaging test phase

.

3.4  Objectives

The objectives of this research are (1) to process X-Ray scan image of the lungs for cancer prediction purpose using digital image processing and (2) to design Adaptive Neuro Fuzzy Inference Systems (ANFIS) for lung cancer early detection in imaging test

3.    Literature Review

Some literatures that have been reviewed for current research purpose are listed below :

Table 1 List of literature review for research purpose

Number	Title	Name	Publisher	Year	Results
1	Cancer Diagnosis Using Modified Fuzzy Network	Essam Al-Daoud	Universal Journal of Computer Science and Engineering Technology	2010	Determining the basic rules to diagnosing cancer using Modified Fuzzy C-Means Radial Basis Functions (MFRBF) and can classified with 97% accuration
2	Lung Cancer Detection Using Image Processing Techniques	Suha  Mohammad AlHabashneh, Sajedah  Emhadi  Al Tarawneh, Weam  Fayg  AlTarawneh, Norah  Shaker  AlJaafreh	Mutah University, Faculty of Engineering Computer Engineering Department	2011	Enchanced X-Ray lung image using Gabor Filter, Fast Fourier Transform, Auto Enhancement Algorithm. Image segmentation, binarization, and masking in X-Ray scan of lung
3	Automated Detection of Early Lung Cancer and Tuberculosis Based on X- Ray Image Analysis	Kim Le	International Conference on signal, speech, and Image Processing WSEAS	2006	Image segmentation approaches for isolating lung in X-ray image and applying small scanning windows to scan pixel that consist cancer or TB nodule
4	Medical Image Classification and Symptoms Detection  using Neuro Fuzzy	Mohd Ariffanan Bin Mohd Basri	Faculty of Electrical Engineering  Universiti Teknologi Malaysia	2008	Applying ANFIS for tumor classification and detection from MRI image of brain through two stage decision making : (1) Principal Component Analysis and (2) ANFIS training using backpropagation method
5	Pengembangan Sistem Kecerdasan Buatan berbasis Adaptive Neuro Fuzzy Inference System untuk Diagnosa Penyakit Kanker Paru	Sylvia Ayu Pradanawati	Jurusan Teknik Fisika-Fakultas Teknologi Industri ITS	2010	Diagnosing lung cancer based on medical history data and X-Ray scan using ANFIS with two and three membership functions