Title page for etd-0411116-205333


[Back to Results | New Search]

URN etd-0411116-205333
Author Chi-Chih Lin
Author's Email Address No Public.
Statistics This thesis had been viewed 5569 times. Download 0 times.
Department Computer Science and Engineering
Year 2015
Semester 2
Degree Master
Type of Document
Language English
Title A Weight-based Approach to the Classification of Uroflowmetry Curves
Date of Defense 2016-06-03
Page Count 97
Keyword
  • Uroflowmetry Curves
  • Classification
  • Uroflow Pattern
  • Flow Index
  • Weighted Grade
  • Abstract Uroflowmetry, which measures the flow and volume of urine during urination, is a common, noninvasive urinary test to diagnose symptoms. The uroflowmetry results can help with assessment of bladder urethra function. The voided volume, the flow curve, and the maximum flow rate are the most frequently mentioned as relevant for interpretation. We can easily get the data of voided volume and maximum flow rate from uroflowmetry, but the flow curve only shows on the uroflowmetry. The result does not show what pattern it belongs to. The physicians need to classify the flow curve by their experiences. However, the results of the flow curve are usually poor agreement among the physicians. To avoid errors in classification of flow curves, before the physician interprets the flow curve, the scales of the horizontal and vertical axes for the output of the flow curve must be adjusted. Moreover, the flow curve may not be the perfect bell-shaped pattern, when the people released the urine from their body, which results in the situation that decided conditions of different patterns may be overlap. In order to increase the degree of the agreement of the results of uroflowmetry curves with the physician’s observation, we survey the parameters in interpreting uroflowmetry, and the uroflow pattern. Therefore, we propose a weight-based approach to the classification of flow curves, which integrates the physician’s interpretation experiences and different weights for different conditions, and the compared pattern with the highest score is the resulting pattern. Besides, we use the coordinates of both axes and several needed parameters to compute the result directly, so that the scale of horizontal and vertical axes are not needed to be adjusted. Moreover, our method not only accurately calculates the area of small curves which will affect the interrupted-shaped pattern, but also correctly computes the rising angle, the drop angle, and the number of drops which are more reliable and effective than the physician’s interpretation. Hence, the results of uroflowmetry efficiently enhance the degree of the agreement among the physicians’ observation by our approach. From our performance study by statistic analysis of comparing the results of our approach with one/two physician’s observation, we show that for the agreement of normal/abnormal types, it is good. Moreover, for the agreement of the specific patterns, it is also good. Although for the agreement of the plateau-shaped and the obstructive-shaped patterns are not good, we find the reason which includes it is also difficult for the physician to define the differences between these two patterns. Therefore, we suggest that our decision rules for pattern classification could be useful in CAI for students with a major in urology.
    Advisory Committee
  • Gen-Huey Chen - chair
  • Chien-I Lee - co-chair
  • Shei-Dei Yang - co-chair
  • You-Chiun Wang - co-chair
  • Ye-In Chang - advisor
  • Files
  • etd-0411116-205333.pdf
  • Indicate in-campus at 5 year and off-campus access at 5 year.
    Date of Submission 2016-06-04

    [Back to Results | New Search]


    Browse | Search All Available ETDs

    If you have more questions or technical problems, please contact eThesys