ขั้นตอนวิธีต้นแบบในการกำหนดสีฟันเทียมด้วยภาพถ่ายดิจิทัล / ธนพงศ์ อินทระ = A prototype algorithm of denture color specification by digital image / Thanapong Intharah
This research proposes an algorithm to specify a denture color based on image analysis. Details of the research are divided into two parts: light problems which highly affect the accuracy of color specification procedure and the algorithm for dealing with the problems. The light problems presented in this research are the instability of incident light and the non-uniform light incidence. The former causes a problem in that the amount of light over an area changes over time. The latter deals with the fact that light illuminate unevenly on different areas. In addition, this research proposes a metric used to illustrate the occurrence of these problems. The proposed algorithm of denture color specification to lessen the light problems is as follows. Firstly, to avoid the problem of instability of incident light, only one image consisting of an unknown color, a set of porcelain reference colors, and a light calibration tool should be taken. Secondly, to lessen non-uniform light incidence, the particular non-uniform light incidence is modeled by Ordinary Kriging using the relative light incidence values measured from the light calibration tool. The model is then use for calibrating the image before color matching with HSV color model is performed. Results from the proposed algorithm are compared with human process as well as the computerized process without light calibration. Experimental results reveal that the proposed algorithm outperforms among all these processes.