| Summary: | 博士 === 國立成功大學 === 電機工程學系 === 89 === Multimedia-oriented Artificial Reproduction Decision Support Systems
Cheng-Huei Yang* and Pau-Choo Chung**
Abstract
Multimedia-oriented artificial reproduction decision support systems use the Internet to transmit, store, process and integrate physical examination findings and historical information of female infertility patients. The systems in general consist of image processing, Internet transmission, database construction, medical information integration and decision analysis. The major advantages of such a system are that it not only improves the efficiency of information access, but also have a pronounced effect on management and integration among treatment related data. In addition, the system provides real time medical diagnostic information for clinical physicians.
Endometrium characteristics of ultrasound images, including thickness and motion frequency, are considered important indexes in infertility treatment. The proposed recursive model in the paper is employed for threshold optimization of motion information segmentation by repeatedly shrinking the dimension of the maximum potential range of motion with an adaptive threshold throughout the whole process. Then, the values of thickness and motion frequency are estimated by using simple mathematics. The clinical database system built is based on the entity-relationship model, which is applied to effectively manage the related infertility data generated in different places and at different time. The data comprises periodic examination findings and image pictures so as to increase the efficiency of access and comparison. The artificial reproduction decision support model is to conclude the potential factors and degree of influence on infertility through a statistical method. Three general regression models including forward selection, backward elimination and the stepwise method are discussed and compared respectively. Thus, the optimized model with deterministic parameters can be obtained and a higher prediction rate of pregnancy can be ensured.
According to the results obtained from the preliminary phase, the developed systems not only led to the creation of novel software, which helps practitioners effectively manage the data and receive the decision information, but also provide a systematic mechanism in dealing with infertility treatment related data.
*The author
**The advisor
|
|---|
