Detect Periodontal status base on OCT

• Main Authors: Jin-Yang Lin, 林進仰
• Other Authors: Yu-Chi Lai
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/e7ye74

碩士 === 國立臺灣科技大學 === 資訊工程系 === 107 === In periodontal diagnosis, it’s very important to know where is teeth, gingiva and the alveolar bone. It can’t be use eye to see. In traditional, there are two ways to measure the periodontal diease. First, the doctor use the peridontl probe to poke in to the periodontal pocket, and to measure the scale of the periodontal probe. This way is fast, but the patient will feel pain. Second, it can be measure in Xray by the distance between Cemento-Enamel Junction and the alveolar bone. Xray cann’t repeat to be used because the radiation. Meanwhile, Optical Coherence Tomography can be used repeatly, because it doesn’t have the radiation and the patient won’t be hurted. This paper did three things. First, we use OCT Image Network to find the tooth information and predict the result with noise OCT result. Second, we modify a term in Snake segmentation to make it consider the information in pervious scan and next scan, and use it to find useful area in a OCT image. Third, we combine the camera calibration and thin-plate spline to project the distance in OCT image to world space distance.