利用無刻度試鏡架建立眼鏡鏡框之尺寸統計

Abstract

眼鏡為許多人不可或缺的重要配件，但由於鏡架尺寸的測量並不受重視，加上眼鏡的購買方式不夠完善，導致許多人會購買到不合尺寸的眼鏡。為了解決這個問題，本論文設計了一個鏡架尺寸量測系統（包含一個基於電腦視覺之辨識技術和對應的特製無刻度試鏡架），讓使用者能夠用在家中自行測量自己合適的鏡架尺寸。使用者先配戴特製試鏡架並將鏡腿長度調整至符合自己的尺寸，之後藉由辨識技術偵測鏡腿上的標記，進而計算鏡腿長度，並將誤差控制於一毫米以內。本論文使用此系統，對大量使用者進行數據的收集，並根據不同使用者的尺寸數據，統計了使用者的鏡腿長度變化幅度及鏡腿曲度對配戴感受之影響，證明了鏡架測量的重要性。

Glasses are inseparable to a lot of people. However, due to the lack of systematic size measurement of the glass frame for shoppers and the inadequacies in the existing practice of business, many people are getting glasses with an inappropriate size. This thesis aims to address such a problem through devising a frame-measuring system (including a recognition system based on computer vision and a corresponding scaleless trial frame), which allows user to perform self- measurement at home. While wearing the trial frame, a user can first adjust at ease the temples of a glass frame to their best-fit lengths. Then, with the marks on the temples detected by the proposed recognition system, the length can be calculated with accuracy within 1mm. In this thesis, such a system is also utilized to collect measured data from a sufficiently large number of users to establish the statistics of the interpersonal variations in their best-fit temple lengths, and to analyze the impacts of different temple curvature on user’s wearing experience. Both suggest the importance of the proposed measuring scheme for glass frames.