Thomas N. Hangartner (Committee Member), Richard J. Sherwood (Committee Member), Julie Skipper (Advisor)
Master of Science in Engineering (MSEgr)
Dental casts have been used extensively to study almost all aspects of the human dentition. These aspects varied from the study of tooth form and morphology, inheritance and genetics, growth and development, occlusion, arch alignment and crowding to mathematical determination of dental arch form.
The aim of this project was to develop a tool to semi-automate the measurement of dental casts that would be precise, accurate and efficient. Measurements include tooth widths, arch lengths and widths, angle of rotation and crown area for each tooth.
The task was divided into two different parts: first, the development of semi-automatic software to analyze 2D dental cast images and implementation of the process in a Graphical User Interface (GUI). The automated sections of the software were to be ideally executed without any user intervention, but it was anticipated that not all of the images would be successfully analyzed. Some factors that affect the automatic analysis are the quality of the casts, variations in tooth shape and image quality. During the analysis, it is possible that some automatically-determined tooth boundaries and arch fits are erroneous. The GUI thus gives an interface for the user to execute the program, view the results of the automated measurements and make any manual adjustments before saving the results of all analyses.
The project objectives were attained. Of a test set of 96 maxilla and mandible images, analysis was successful on all of the images with none or limited manual intervention.
To assess accuracy, the results obtained using the software were successfully compared to those using traditional manual techniques taken with calipers, protractors, scales, thread, etc. The percentage error for measurements obtained using calipers were less than 1%. The developed software tool provides results that are more accurate and precise than those from manual analyses. The automated analysis process is also more effective than manual image analysis in that not all measurements can be obtained manually; the program automatically generates an output file containing over 260 variables of interest. Intraoperator and interoperator error analysis was also performed. We showed that the mean percentage intraoperator errors for the mesio-distal distances were 3.46% for the maxilla and 3.48% for the mandible and those for the bucco-lingual distances were 3.29% for the maxilla and 2.97% for the mandible. The mean percentage interoperator errors for mesio-distal distances were 5.82% for the maxilla and 4.46% for the mandible, and those for the bucco-lingual distances were 3.32% for the maxilla and 3.81% for the mandible. The software is currently being used to analyze over 2600 images.
Department or Program
Department of Biomedical, Industrial & Human Factors Engineering
Year Degree Awarded
Copyright 2011, all rights reserved. This open access ETD is published by Wright State University and OhioLINK.