01-09-2012, 10:58 AM
Computerized Medical Diagnosis
[attachment=33066]
INTRODUCTION
The use of computer in medical applications has been widely increased dramatically over the last three decades. Traditionally, in this field, computers have been used to read, scan, store, process, retrieve and analyze medical information. They have also been used as a tool in the diagnosis and care of patients.
More recently, computerized image processing techniques have been used to improve the picture quality and images can be analyzed to highlighten areas of interest or to extract meaningful diagnostic features that can provide objective evidence to aid the human decision making process. The integration of the two types of systems – databases and computers–aided feature extraction and allowed diagnostic information to be attached to patient records. This information can, then, be easily searched and cross referenced.
An expert system is a computer program that has the knowledge of an expert system built into a series of rules. The expert system then finds the solution to a problem, such as the diagnosis of a medical disease.
Machine learning involves getting computer to induce rules, or statistical models, from a set of "training" examples. These examples from the problem domain can either have the correct class labels associated with them, often called supervised learning, or sub–classes of examples can automatically be found from within the data, often called unsupervised learning. Examples of machine learning techniques are artificial neural networks and decision trees.
The Diagnosis of Brain Tumors:
The occurrence of brain tumors is one of the leading causes of cancer death in young adults and the second fastest growing cause of cancer death among people over age 65. Persons with a brain tumor often have symptoms like gradual loss of movement or sensation in an arm of leg, loss of vision or speech difficulty. In hospital, a radiological investigation may demonstrate an abnormality in the brain. Since the symptoms of the patient can easily be confused with other disorders of the brain, e.g. infract or absence, the first step is to identify the disorder as being a tumor. If a brain tumor has been diagnosed, the next very important step is to identify the type of tumor. The tumor type is dependent on the type of cell the tumor originates from. In this thesis, the focus is on two types of tumors: astrocytoma and meningioma. The first tumor types arise from the brains supportive tissue, and are collectively called gliomas. Meningiomas are tumors that arise from the leptomeninges (tissue that serves as the lining of the brain). The tumor grade indicates the level of tumor malignancy. Tumors are graded on their growth rate, vascularity (blood supply), presence of a necrotic center, invasive potential (border distinctness) and similarity to normal cells. Malignant tumors may contain several grades of cells. The most malignant grade of cell found determines the grade for the entire tumor, even if most of the tumor is a lower grade[3].
Tumor Segmentation:
Automated detection of tumors in different medical images is motivated by the necessity of a high accuracy when dealing with a human life. Also, the computer assistance is demanded in medical institutions due to the fact that it could improve the results of humans in such a domain where the false negative cases must be at a very low rate. It has been proved that a double reading of medical images could lead to a better tumor detection. But the cost implied in a double reading is very high, that’s why good software to assist humans in medical institutions is of great interest nowadays.
Different approaches are needed as function of the medical images that must be studied. The technique that produced those images is very important in order to know what to apply to a certain medical image in order to get better results. A lot of methods have been proposed in the literature for CT (Computed Tomography) scans, different types of
X-rays, MRI images and other radiological techniques. With all, this effort which has been done in the research field, there is a lot of place for improvements and the medical image processing is a domain in continuous expansion. Why is this domain in continuous expansion and there are no good accepted methods? This is due to the fact that in such an important domain, the accuracy must be very high and the false negative rate must be low. The problem is that is not very easy to obtain such results. Anyway, the idea is to reduce as much as possible the human errors by assisting the physicians and the radiologists with some software that could lead to better results. This is important for the human life.
Genetic Algorithm:
Genetic algorithm is a new kind of theory and way in the pattern recognition field and it is a kind of bionic technology in which simulation mechanic is concerning with the production and evolution processing of all life and intellectual. Its main features is to adopt the colony searching game and the information exchanging among the individuals instead of depending on the gradient problems. So, it is very suited to be used in the problems that are difficult to be solved when using the traditional searching method. It can also be used in the combination optimization, machine learning, self–adaptive control, programming design and man made life field, etc. Accordingly, it is the key technology used in the intellectual computer in the 21st century[17]. It is a technique for searching the fitness landscape for a highly fit individual while a fitness landscape search space is formed from the representation of all possible genotypes and their fitness value[18].
Neural Network:
Artificial neural networks (ANN) are computational paradigms based on mathematical models, that unlike traditional computing, have a structure and an operation that resembles that of the mammal brain. Artificial neural networks or neural networks for short, are also called connectionist systems, parallel distributed systems or adaptive systems, because they are composed by a series of interconnected processing elements that operate in parallel. Neural networks lack centralized control in the classical sense, since all the interconnected processing elements change or “adapt” simultaneously with the flow of information and adaptive rules[20].
One of the original aims of the artificial neural networks (ANN) was to understand and shape the functional characteristics and computational properties of the brain when it performs cognitive processes such as sensorial perception, concept categorization, concept association and learning. However, today a great deal of effort is focused on the development of the neural networks for the applications such as pattern recognition and classification, data compression and optimisation[20].
Architecture of ANNs:
A generic artificial neural network can be defined as a computational system consisting of a set of highly interconnected processing elements, called neurons, which process information as a response to external stimuli. An artificial neuron is a simplistic representation that emulates the signal integration and threshold firing behaviour of biological neurons by means of mathematical equations. Like their biological counterpart, artificial neurons are bound together by connections that determine the flow of information between peer neurons. Stimuli are transmitted from one processing element to another via synapses or interconnections, which can be excitatory or inhibitory. If the input to a neuron is excitatory, it is more likely that this neuron will transmit an excitatory signal to the other neurons connected to it. Whereas, an inhibitory input will most likely be propagated as inhibitory[20].
Neural Networks in Medicine:
Artificial neural networks (ANN) are currently a hot research area in medicine and it is believed that they will receive extensive application to biomedical systems in the next few years. At the moment, the research is mostly on modeling parts of the human body and recognizing diseases from various scans (e.g. cardiograms, CT scans, ultrasonic scans, MRI, PET).
Neural networks are ideal in recognizing diseases by using scan since there is no need to provide a specific algorithm on how to identify the disease. Neural networks are learned by an example so the details of how to recognize the disease, are not needed. What is needed, is a set of examples that are representative of all the variations of the disease. The quantity of examples is not as important as the quality. The examples need to be selected very carefully, if the system is to perform reliably and efficiently[22].
Bioinformatics:
Bioinformatics is a new scientific discipline that combines biology, computer science, mathematics, and statistics into a broad-based field that will have profound impacts on all fields of biology. Bioinformatics is expected to substantially impact on scientific, engineering and economic development of the world.
It is the study of biological systems by using computational techniques. It represents a relatively new area of computer science and is of an increasing relevance to science as large amounts of data are generated by technologies which are designed for measuring biological systems[22].
Research and development in bioinformatics and computational biology require the cooperation of specialists from the fields of biology, computer science, mathematics, statistics, physics, and such related sciences. It is the comprehensive application of mathematics (e.g., probability and graph theory), statistics, science (e.g., biochemistry), and computer science (e.g., computer algorithms and machine learning) to the
understanding of living systems. Bioinformatics is fast emerging as an important discipline for the academic research and industrial application.
[attachment=33066]
INTRODUCTION
The use of computer in medical applications has been widely increased dramatically over the last three decades. Traditionally, in this field, computers have been used to read, scan, store, process, retrieve and analyze medical information. They have also been used as a tool in the diagnosis and care of patients.
More recently, computerized image processing techniques have been used to improve the picture quality and images can be analyzed to highlighten areas of interest or to extract meaningful diagnostic features that can provide objective evidence to aid the human decision making process. The integration of the two types of systems – databases and computers–aided feature extraction and allowed diagnostic information to be attached to patient records. This information can, then, be easily searched and cross referenced.
An expert system is a computer program that has the knowledge of an expert system built into a series of rules. The expert system then finds the solution to a problem, such as the diagnosis of a medical disease.
Machine learning involves getting computer to induce rules, or statistical models, from a set of "training" examples. These examples from the problem domain can either have the correct class labels associated with them, often called supervised learning, or sub–classes of examples can automatically be found from within the data, often called unsupervised learning. Examples of machine learning techniques are artificial neural networks and decision trees.
The Diagnosis of Brain Tumors:
The occurrence of brain tumors is one of the leading causes of cancer death in young adults and the second fastest growing cause of cancer death among people over age 65. Persons with a brain tumor often have symptoms like gradual loss of movement or sensation in an arm of leg, loss of vision or speech difficulty. In hospital, a radiological investigation may demonstrate an abnormality in the brain. Since the symptoms of the patient can easily be confused with other disorders of the brain, e.g. infract or absence, the first step is to identify the disorder as being a tumor. If a brain tumor has been diagnosed, the next very important step is to identify the type of tumor. The tumor type is dependent on the type of cell the tumor originates from. In this thesis, the focus is on two types of tumors: astrocytoma and meningioma. The first tumor types arise from the brains supportive tissue, and are collectively called gliomas. Meningiomas are tumors that arise from the leptomeninges (tissue that serves as the lining of the brain). The tumor grade indicates the level of tumor malignancy. Tumors are graded on their growth rate, vascularity (blood supply), presence of a necrotic center, invasive potential (border distinctness) and similarity to normal cells. Malignant tumors may contain several grades of cells. The most malignant grade of cell found determines the grade for the entire tumor, even if most of the tumor is a lower grade[3].
Tumor Segmentation:
Automated detection of tumors in different medical images is motivated by the necessity of a high accuracy when dealing with a human life. Also, the computer assistance is demanded in medical institutions due to the fact that it could improve the results of humans in such a domain where the false negative cases must be at a very low rate. It has been proved that a double reading of medical images could lead to a better tumor detection. But the cost implied in a double reading is very high, that’s why good software to assist humans in medical institutions is of great interest nowadays.
Different approaches are needed as function of the medical images that must be studied. The technique that produced those images is very important in order to know what to apply to a certain medical image in order to get better results. A lot of methods have been proposed in the literature for CT (Computed Tomography) scans, different types of
X-rays, MRI images and other radiological techniques. With all, this effort which has been done in the research field, there is a lot of place for improvements and the medical image processing is a domain in continuous expansion. Why is this domain in continuous expansion and there are no good accepted methods? This is due to the fact that in such an important domain, the accuracy must be very high and the false negative rate must be low. The problem is that is not very easy to obtain such results. Anyway, the idea is to reduce as much as possible the human errors by assisting the physicians and the radiologists with some software that could lead to better results. This is important for the human life.
Genetic Algorithm:
Genetic algorithm is a new kind of theory and way in the pattern recognition field and it is a kind of bionic technology in which simulation mechanic is concerning with the production and evolution processing of all life and intellectual. Its main features is to adopt the colony searching game and the information exchanging among the individuals instead of depending on the gradient problems. So, it is very suited to be used in the problems that are difficult to be solved when using the traditional searching method. It can also be used in the combination optimization, machine learning, self–adaptive control, programming design and man made life field, etc. Accordingly, it is the key technology used in the intellectual computer in the 21st century[17]. It is a technique for searching the fitness landscape for a highly fit individual while a fitness landscape search space is formed from the representation of all possible genotypes and their fitness value[18].
Neural Network:
Artificial neural networks (ANN) are computational paradigms based on mathematical models, that unlike traditional computing, have a structure and an operation that resembles that of the mammal brain. Artificial neural networks or neural networks for short, are also called connectionist systems, parallel distributed systems or adaptive systems, because they are composed by a series of interconnected processing elements that operate in parallel. Neural networks lack centralized control in the classical sense, since all the interconnected processing elements change or “adapt” simultaneously with the flow of information and adaptive rules[20].
One of the original aims of the artificial neural networks (ANN) was to understand and shape the functional characteristics and computational properties of the brain when it performs cognitive processes such as sensorial perception, concept categorization, concept association and learning. However, today a great deal of effort is focused on the development of the neural networks for the applications such as pattern recognition and classification, data compression and optimisation[20].
Architecture of ANNs:
A generic artificial neural network can be defined as a computational system consisting of a set of highly interconnected processing elements, called neurons, which process information as a response to external stimuli. An artificial neuron is a simplistic representation that emulates the signal integration and threshold firing behaviour of biological neurons by means of mathematical equations. Like their biological counterpart, artificial neurons are bound together by connections that determine the flow of information between peer neurons. Stimuli are transmitted from one processing element to another via synapses or interconnections, which can be excitatory or inhibitory. If the input to a neuron is excitatory, it is more likely that this neuron will transmit an excitatory signal to the other neurons connected to it. Whereas, an inhibitory input will most likely be propagated as inhibitory[20].
Neural Networks in Medicine:
Artificial neural networks (ANN) are currently a hot research area in medicine and it is believed that they will receive extensive application to biomedical systems in the next few years. At the moment, the research is mostly on modeling parts of the human body and recognizing diseases from various scans (e.g. cardiograms, CT scans, ultrasonic scans, MRI, PET).
Neural networks are ideal in recognizing diseases by using scan since there is no need to provide a specific algorithm on how to identify the disease. Neural networks are learned by an example so the details of how to recognize the disease, are not needed. What is needed, is a set of examples that are representative of all the variations of the disease. The quantity of examples is not as important as the quality. The examples need to be selected very carefully, if the system is to perform reliably and efficiently[22].
Bioinformatics:
Bioinformatics is a new scientific discipline that combines biology, computer science, mathematics, and statistics into a broad-based field that will have profound impacts on all fields of biology. Bioinformatics is expected to substantially impact on scientific, engineering and economic development of the world.
It is the study of biological systems by using computational techniques. It represents a relatively new area of computer science and is of an increasing relevance to science as large amounts of data are generated by technologies which are designed for measuring biological systems[22].
Research and development in bioinformatics and computational biology require the cooperation of specialists from the fields of biology, computer science, mathematics, statistics, physics, and such related sciences. It is the comprehensive application of mathematics (e.g., probability and graph theory), statistics, science (e.g., biochemistry), and computer science (e.g., computer algorithms and machine learning) to the
understanding of living systems. Bioinformatics is fast emerging as an important discipline for the academic research and industrial application.