Research Of Intelligent Diagnosis And Treatment System For Fracture Based On Hybrid Reasoning

Intelligent diagnosis and treatment system for fractures is an important extension of medical robot for fracture resetting. X-ray image provides the key foundation for diagnosis and treatment, and medical record provides assistant basis for the system. This system comprises knowledge in many fields, such as computer technique, modern control theory, digital image processing theory, pattern recognition technique, artificial intelligence and expert system theory, sensor technique, biomechanics, anthropotomy, fracture diaplasis theory and bare-handed fracture resetting theory. The system is a comprehensive medical system being able to draw a diagnostic conclusion of human fractures and thus decide treatment scheme and get it implemented. The system is composed of two sub-systems, diagnosis sub-system and treatment sub-system. The diagnosis sub-system, as a separate medical expert system, can help experts of this field to complete fracture diagnosis, that is, to give a clinical deterministic diagnostic conclusion and treatment scheme. Treatment sub-system, as the important integrant of medical robot for fracture resetting, can implement fracture treatment scheme after diagnosis by extracting features of resetting, calculating parameters of resetting, editing programmes of resetting, and driving control system and machine structure to directly carry out resetting operation on patients. The two sub-systems can be utilized separately and dependently to realize sequential link between diagnosis and treatment. Code diagnosis is the crucial element in the process of achieving clinical deterministic conclusion. It is advanced that clinical experience and knowledge can be translated into some rules by introducing rule-based reasoning theory of intelligent diagnostic technique. The decisive diagnostic features can be extracted by means of man-machine intersection to get the crytic information in X-ray images of fracture patients and then diagnostic code of fracture can be presented. Clinically, correlative information in medical record must be combined into the procedure of achieving deterministic diagnostic conclusion. It is presented in this dissertation that case-based reasoning technique can simulate thinking mode and diagnostic train of thought of experts to implement intelligent diagnosis of fractures. Fuzzy set is put forward to represent medical record cases including many ambiguous concepts; a hierarchical structure is advanced to organize the cases and to realize efficient retrieval of cases. Calculation method of similarity between fuzzy sets is introduced in the process of similarity between cases in order to implement the matching of similar cases. If the most similar case is not able to be searched in casebase to match the current patient or diagnostic conclusion need to be corrected, the technique presented here based on the combination of case-based reasoning and rule-based reasoning can not only make up for the deficiency of case-based reasoning used alone, but also improve ability to find the solution to the system. Quantizing and parametrizing of operation for fracture resetting is the key to implement therapy in treatment sub-system. It has long been a difficult problem how to transform sense and experience of bare-handed resetting of traditional Chinese medical science into definite resetting operation in the procedure of realizing automatic resetting of fracture. In the light of peculiarities of femoral-neck fractures, usually happening to old people, it is advanced that resetting of femoral-neck can be quantized into three actions, including pull, left/right shift and inside rotation, whose features can be reflected in the X-ray images, including nine points, six lines, two angles and one distance. Proper image processing and analyzing strategy are brought forward according to the characteristics of X-ray images captured by C-shape-arm X-ray machine. At the same time, master-slave coordinate method is brought forward in order to calculate feature angles and feature distances on the basis of analyzing specialties of diaplasis parameters and diaplasis demands. The method provides a novel reference for calculation of feature parameters and predigests calculation steps of these parameters. Four-group-sixteen-frame algorithm is put forward to solve the problem of being unable to calculate precisely the stress value under treatment of lingering concrescence and non-union of fractures.