運用二維相片組構三維臉像

Abstract

本篇論文提出一套簡潔的三維臉像建構模式，能夠在電腦上模擬實際 的三維 臉像，並可透過由電腦控制的自動化機具雕刻出實際成品。本模 式運用三張-右側 、正面、左側同時取像的照片，以最佳化的方式分別計 算出左、右臉的控制點的 三維座標，再模擬出三維的臉像。所需之設備 為一般的個人電腦及常見軟體，不 需任何精密或昂貴的儀器，甚適合一 般應用。研究結果顯示本模式能有效的展現 立體臉像並可模擬不同的表 情。 This research proposes an efficient model to formulate the 3D image of a specific person's face based on three 2D images. We can not only simulate a specific person's face with various emotions but also sculpture the face by automatic machinery (ex: DYNA2200). Firstly, we take three pictures from right, front and left face. The needed equipment are three general cameras, a personal computer and common software. No expensive or precision instrument is required, so this method is suitable for many applications. Outcome of this study reveals that this model can profile 3D faces and simulate different facial expressions effectively.