Diagnostic System of the Maxillofacial Area Diseases (Dentistry): based on Neural Networks, Artificial Intelligent methods and Expert Systems.
Serge G. Manukov1, 2, George M. Papoudurakis1, George G. Gogichaishvili2,
1. Technological Educational Institute of Crete, Department of Applied Informatics and Multimedia, TEI of Crete, P.O. Box 71500, Estavromenos, Iraklion - Crete, GREECE tel: +(30 2810) 37-98-02, E-mail: firstname.lastname@example.org
Stomatology is a field of clinical medicine, which studies symptomatology, diagnostics and treatment of maxillofacial area diseases and on the basis of system approach, develops diagnostics methods of physiological and pathological processes, proceeding in functional systems. Significant achievements in the study of etymology and pathogenesis of oral region disorders gives a stomatologist an opportunity to make a more correct diagnose and carry out the right treatment.
Not matter how well the dentist will be prepared his practical activity depends on the equipment quality in the clinic. The modernization of material resources in stomatological institutions and dental laboratories would provide them all necessary equipment to execute medical assistance on the highest level.
Determination of disease proceeding and symptoms for the given patient, his physical and psychical state, level and type of morphological and functional disorders, is possible only at correct and careful clinical research. Detection of etiological moments and pathogenesis assists in the diagnosis determination. Diagnosis – is a brief written medical conclusion of the existing disease (trauma), described in medical terms, denoting the name of disease (trauma) - its form, and defining individual peculiarities of the patient’s organism. The study of diseases determination methods is called «diagnostics» and is an important part of any medical specialty. Medical diagnostics is based on different methods of research and diseases determination, their severity and patient’s state of organism with the purpose to select and carry out necessary treatment and effective prophylactic measures, preventing the development of complications and recurring disease. Diagnostics is a complex cognitive process. To carry out the correct diagnostic process it is necessary to study and be able to apply on practice several tests. It is necessary to know principal and specific symptoms of the oral region disorder diseases and their classification. Advanced medical thinking, knowledge of the analysis and synthesis of established subjective and objective symptoms, logically proved methods of laboratory investigations allows in defining and specifying disease’s etiology and its pathogenesis. Due to this, without the knowledge of physiological standards and possible physiological variants of separate organ’s functioning, forming dental and oral system, their topographic and functional interrelations, it is impossible to master semiology in detail, the whole diagnostics process, and therefore, correctly determine the diagnosis and accomplish all necessary treatment procedures.
Making out a diagnosis is connected with several issues: people are used to get tired and forget, they may also make careless mistakes or make hasty choice not taking into account an important symptom, which on the first sight may be feebly marked or does not exceed the value of the most vividly proceeding, but not the main symptoms. Purely classic manifestations of diseases occur seldom, more frequently there are various deflections or combinations of diseases, and as a result changes classic scheme of clinical course. An expert system is proposed contains the knowledge, obtained by experienced dentists, and information from relative literature and tutorials. In the form of such advising system the stomatologist will get expert assistance containing any difficult diagnostic situation. The structure of the proposed expert system is illustrated in Figure 1.
Initially, the dentist gathers information from the patient in the form of complaints (toothache or the mouth doesn’t open), and also medical information based on patient’s examination (reddening mucous tunic in area of 16th tooth), the dentist provides this information as symptoms and their values through an interface engine. Then the analyzer takes the information provided by the interface engine and performs feature extraction on the various data in order to provide to the expert system database a set of symptoms.
The knowledge base system provides diagnosis corresponding for the inserted symptoms. Afterwards the expert system processor recalculates the confidence coefficients per each provided diagnosis based on the initial set of symptoms.
The expert system processor chooses the symptom; which has maximum dispersion in the confidence coefficient. After checking up on no coincidence and if there is not such symptom-value in the analyzer output, it displays a question for the dentist in the form of symptom and different values. After the dentist receives the question he clarifies it with the patient. If additional tests are required they are performed and inputted to the expert system. These tests could be x-rays, electroodontometry, common blood test, biopsy, gum analyses and etc. The analyzer determines all the symptoms it can extract from the provided data. Then the expert system processor recalculates the confidence coefficient and updates information in the database. At this stage the expert system discards the diagnosis with low confidence coefficients. The remaining set of diagnoses and its confidence coefficients are associated with every possible symptom of the knowledge base and if the expert system finds a diagnosis with a high confidence coefficient it provides the diagnosis otherwise, the expert system selects the symptom with the maximum dispersion and consults the dentist and this procedure is repeated until the expert system finds a diagnosis with a high confidence coefficient
An important and very complex part of the expert system is the analyzer. Its purpose is to analyze the input data and provide a set of symptoms to the expert system database. It is a hybrid system containing, among other things, several neural networks and sub expert systems. The functionality of the analyzer can be demonstrated by describing the procedure performed on an x-ray input image and is shown in Figure 2. Initially the x-ray data (scanned or automatically received from digital x-ray sensor) is preprocessed using digital image processing techniques in order to improve its quality. Then feature extraction is performed in order to obtain valuable image impacts such as tissues and artificial impurities. A neural network will processes the feature extraction data in order to identify the objects. Finally, the sub expert system will calculate all necessary lengths and provides set of all detected symptoms as outputs.
The final version of the paper will provide tests results and conclusions of the proposal expert system
1. Greenwell, Mike, 1988 - Knowledge engineering for
expert systems. Horwood,
2. Masters, Timothy, 1994 - Signal and image processing
with neural networks: a C++ sourcebook. Wiley.
3. Cleal, David, 1988 - Knowledge based systems: Implications
for human-computer interfaces. Horwood.
4. Papadourakis, George, 2001 - Neural networks and expert
systems in medicine and healthcare (Proceedings of the Fourth International
Conference). TEI of
First Published March 2003