Research On Digital Detection And Analysis System Of Dental Occlusion And Its Application

Occlusion is the phenomenon of extensive contact between the maxillary and mandibular dentition,which is the basis for the oral cavity to achieve masticatory function and runs through the whole process of prevention,diagnosis,and treatment in dentistry.The occlusal contact detection methods mainly used in clinical practice are subjective-dependent and have limited sensitivity and accuracy.Quantitative detection methods represented by digital occlusal analysis systems have become an urgent need for dentists,patients,and researchers.Foreign commercial products that can achieve quantitative occlusal detection occupy the market by their technical advantages,are expensive,and have relatively obsolete analysis and evaluation functions;the lack of literature and research on the design of digital occlusal analysis systems further limits the popular application of quantitative detection methods in the clinical practice.In recent years,the state and provinces have issued documents to explicitly support the domestic substitution of medical equipment.To develop a digital detection and analysis system for oral occlusion,this dissertation conducts research on key technologies and methods of occlusion detection,such as dental position partitioning methods,analysis and calculation of occlusion quantitative assessment indexes,based on which a digital detection and analysis system for oral occlusion was designed and developed.The main work of this dissertation is as follows.1.To align the occlusal force distribution with the actual dental position,a dental position partitioning method based on a three-dimensional dental arch model was established,which mainly includes three-dimensional dental arch model segmentation,tooth position area extraction,and dental position area numbering methods.For the problem of tooth position partitioning and numbering in the case of tooth loss,a tooth position area numbering method was improved by combining numerical analysis and the tooth position partitioning method,which solved the problem that the existing method could not complete the tooth position numbering.The correct rate of this method in various types of tooth loss cases was tested using a public data set,and all of them are above 91.67%.2.Relevant clinical reports were synthesized and several quantitative evaluation indexes were extracted,the degree of variation of each index in normal occlusion and malocclusion populations were compared by Meta-analysis,and six quantitative occlusion evaluation indexes with significant differences were selected by this method,and applied to the digital detection and analysis system of dental occlusion designed and developed in this dissertation.The sensitivity and bias analysis of the Meta-analysis results were conducted to verify the reliability of the selected results,and the calculation method of each selected index was designed and implemented.3.The system requirements were comprehensively analyzed.Based on the existing hardware,a digital detection and analysis system of dental occlusion was designed and implemented,including structural design and software design.To test whether the system can meet the requirements of clinical application,this dissertation first summarized the error sources of the system in clinical application,clarified the test contents,collected the occlusal data of several subjects with the partner hospital as the analysis samples,and conducted the tests to verify the system repeatability,sensitivity of changing sensors,and the consistency of measurement between the system in this dissertation and the commercial product.The results show that the measurement results of this system are consistent with the commercial product and can meet the requirements of clinical use.4.Angle’s classification based on the digital detection and analysis system of dental occlusion developed in this dissertation was investigated.The occlusion quantitative evaluation indexes preferred in this dissertation were used as the feature input to construct Angle’s classification model based on the occlusal pressure distribution,and the performance of each model was evaluated and compared.To address the inherent data imbalance problem among the various categories of Angle’s classification,the ADAS YN method based on oversampling was used for processing,which improved the model classification performance.The correct rate of the random forest model after unbalanced processed is 87.83%,which is better than the performance indexes of similar models in foreign literature,further demonstrating the application value of this system in dentistry aided diagnosis.