以三維主動外觀模型在平面影像上定位立體人臉特徵

3D Active Appearance Models for Aligning Faces in 2D Images

陳俊維(Chun-Wei Chen)

指導教授 ： 王傑智

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2008年)

https://doi.org/10.6342/NTU.2008.01449

摘要

近年來，人與機器人互動成為機器人學及相關領域中熱烈討論的一個重要課題。由於人臉可能是人類最富有表現自己想法的部位，因此，觀察人臉成為了人與機器人互動中最重要的任務之一。過去二十年，主動外觀模型（Active Appearance Model）成功地建立了人臉模型並且能在影像中找到人臉細部特徵的位置，如眼睛、鼻子和嘴巴。主動外觀模型可成為分析與瞭解人類表情的重要工具之一。但現有主動外觀模型利用平面形狀模型來為立體的人臉建模，其中立體的資訊就被忽略了。因此，當像機拍攝到非正臉的人臉時，就有極高可能無法正確地找到這些特徵。在此篇論文中，我們提出三維立體主動外觀模型來克服此問題。我們所提出之三維立體主動外觀模型包含了一立體形狀模型（3D Shape Model）以及外觀模型（Appearance Model）。此模型於訓練階段時僅需要正臉的人臉資料即可。即使在有轉動姿勢的人臉的情況下，也可成功地定位人臉的細部特徵。經由20個人的資料庫所訓練並測試的實驗結果顯示出所提出的演算法的成功辨識率更可達80%。

關鍵字

主動外觀模型 ； 立體主動外觀模型 ； 電腦視覺 ； 人機互動

並列摘要

Perceiving human faces is one of the most important functions for human robot interaction. The active appearance model (AAM) is a statistical approach that models the shape and texture of a target object. According to a number of the existing works, AAM has a great success in modeling human faces. Unfortunately, the traditional AAM framework could fail when the face pose changes as only 2D information is used to model a 3D object. To overcome this limitation, we propose a 3D AAM framework in which a 3D shape model and an appearance model are used to model human faces. Instead of choosing a proper weighting constant to balance the contributions from appearance similarity and the constraint on consistent 2D shape with 3D shape in the existing work, our approach directly matches 2D visual faces with the 3D shape model. No balancing weighting between 2D shape and 3D shape is needed. In addition, only frontal faces are needed for training and non-frontal faces can be aligned successfully. The experimental results with 20 subjects demonstrate the effectiveness of the proposed approach.

並列關鍵字

Active Appearance Model ； 3D Active Appearance Model ； Computer Vision ； Human Robot Interaction

參考文獻

Baker, S. & Matthews, I. (2004). Lucas-kanade 20 years on: A unifying framework. International Journal of Computer Vision, 56(3), 221 – 255.

Cootes, T. & Taylor, C. (1992). Active shape models- - smart snakes. British Machine Vision Conference.

Cootes, T. & Taylor, C. (1994). Modelling object appearance using the greylevel surface. In Proceedings of the British Machine Vision Conference.

Cootes, T. & Taylor, C. (2001). Constrained active appearance models. Computer Vision, 2001. ICCV 2001. Proceedings. Eighth IEEE International Conference on, 1, 748–754 vol.1.

Cootes, T. F., Cooper, D. H., Taylor, C. J., & Graham., J. (1991). A trainable method of parametric shape description. British Machine Vison Conference, 54–61.