跳到主要內容

臺灣博碩士論文加值系統

(44.192.48.196) 您好!臺灣時間:2024/06/14 18:36
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:蔡孟翰
研究生(外文):Meng-Han Tsai
論文名稱:使用彈性物體轉換完成之臉部表情合成系統研究
論文名稱(外文):Facial expressions synthesizing system using Elastic Body Spline transform
指導教授:郭鐘榮
指導教授(外文):Chung J. Kuo
學位類別:碩士
校院名稱:國立中正大學
系所名稱:電機工程研究所
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2000
畢業學年度:88
語文別:中文
論文頁數:43
中文關鍵詞:臉部模型原始臉部模型有表情之臉部模型人體測量學特徵點非特徵點
外文關鍵詞:Facial modelOriginal face modelExpressional face modelElastic body splineanthropometryFeature pointsNon-feature points
相關次數:
  • 被引用被引用:1
  • 點閱點閱:162
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
在本篇論文中,我們發展了一個臉部表情合成系統,這一個系統利用了所謂的Elastic body spline (EBS)之方法,合成二維的臉部表情之影像。Elastic body spline是一種座標轉換的技術,它對物體建立了一種物理模型,在這個物理模型中的物體都是3-D或2-D之homogeneous, isotropic的elastic body。因為臉部表情的產生是由臉部肌肉的變形而來,因此我們可以將人類臉部視為一種elastic body去進行處理。
在我們的臉部表情合成系統中,首先我們需要建立一個2-D的臉部模型,測量人臉上的feature points位置。接著根據EBS以及人體測量學(anthropometry)的原理,我們可以將僅有肌肉變動,因情緒產生之臉部表情;以及因說話發聲如數字,母音,雙母音以及子音而產生的臉部表情等等臉部表情合成出來。若臉部有轉動的現象,我們仍可以使用我們的系統來計算出包含表情與頭部轉動後的2-D臉部模型。也就是說我們的系統可以透過較為簡單的計算,直接近似出在3-D世界中人類臉部轉動以及變形兩者結合所產生的效應。另外,根據人體測量學的原理,臉部每一個區域的變形都是不同的。因此在我們的系統之中,我們可以利用不同臉部參數來對不同的臉部區域產生適當的變形。
在前人的工作中,臉部動畫的模擬所使用的方法有所謂的Action units以及其他線性組合的方式,來建立出臉部表情,如此一來並沒有考慮到臉部轉動以及臉部肌肉力量的連續特性。在此,我們將上述的考量在我們設計系統時包含進來,因此應能得到較佳結果。最後因為我們的feature points是依照MPEG-4標準來定義的,因此我們的系統可以在MPEG-4相關應用中使用。
In this paper, we develope a face expressional model synthesizing system that can synthesize two-dimensional (2-D) facial expressions by a precise method called elastic body spline (EBS). EBS is a coordinate transformation based on a physical model of homogeneous, isotropic three-dimensional and two-dimensional elastic body. Because the facial expressions are generated as the face muscles perform deformations, we can thus model the human''s face as an elastic body.
In our face expressional model synthesizing system, we need to construct a 2-D facial model to measure the feature points on the subject face. Based on EBS and anthropometry, we can completely synthesize the facial expressions such as simple and complex muscle movement and the pronunciation of numbers, vowels, diphthongs, and consonants. If the subject''s face rotates, we still can use our system to calculate 2-D facial model with expressions and rotations. That is, we can directly use our system to approximate the combined effect of rotation and deformation of 3-D human face that can simplified the computation. Furthermore, according to the anthropometry, the deformation of every region on the face is different. In our system, we can use different parameters to generate suitable deformation.
Most of the past works in facial animation used action units and their linear combination to construct facial expressions that didn''t consult the face rotation and the force field connectivity of face''s muscle. Here, we consider these issues in our design such that we can get better results. Because our feature points are defined based on MPEG-4 standard, our system can be used for MPEG-4 related applications.
Chapter 1
Introduction2
Chapter 2
Principle of Volume spline method and Thin-plate splines6
2.1 Volume splines method6
2.2 Thin-plate splines6
2.3 Development of Thin-plate splines8
2.4 Solution of Thin-plate splines10
Chapter 3
Principle of Elastic Body Spline transform6
3.1 Navier partial differential equations PDEs6
3.2 Development of Elastic body spline transform8
3.3 Solution of Elastic body spline transform10
Chapter 4
Facial expressions synthesizing system using Elastic body spline transform 12
4.1 Development of system12
4.1.1 Create original Face model13
4.1.2 Face region selection and demultiplexer15
4.1.3 2D Inverse elastic body spline transform system18
4.1.4 2D Elastic body spline transform system19
4.1.5 Region boundary reprocess using 2D Elastic body spline
transform20
Chater 5
Simulation results and Conclusion23
5.1 2D generic model23
5.2 2D original face model 24
5.3 2D expressional face model 26
5.3.1 2D expressional face model (/u/)26
5.3.2 2D expressional face model (/I/)28
5.4 Conclusion 26
[1] R. Bajcsy and S. Kovacic, "Multiresolution elastic matching ," Computer vision, Graphic, and Image Processing, vol.46(1), Apr.,1989, pp.1-21.
[2] I. S. Sokolnikoff, Mathematical Theory of Elasticity, McGraw-Hill, New York, 1956.
[3] S. Timoshenko and J. N. Goodier, Theory of Elasticity, McGraw-Hill, New York, 1951.
[4] M. H. Davis, A. Khotanzad, D. P. Flamig and S. E. Harms, "A physics-based coordinate transformation for 3D image matching ," Medical Imaging, IEEE Transactions, vol. 16, no. 3, pp. 317 -328, June 1997.
[5] R. L. Harder and R. N. Desmarais, "Interpolation using surface splines," J. Aircraft, vol. 9, no. 10, pp. 189-191, 1972.
[6] C. J. Kuo, R. S. Huang, and T. G. Lin, ŗD facial model estimation from single front-view facial image," IEEE Transactions on Circuits and Systems for Video Technology.
[7] John W. Hole, Karen A. Koos, 人體解剖學, 1992.
[8] "MPEG-4 Committee Draft," Coding of moving pictures and audio, ISO/IEC/JTC1/SC29/WG11 N2196, Mar. 1998.
[9] F. L. Bookstein, "Principal warps: Thin-plate splines and decomposition of deformations," IEEE trans. Pattern Anal. Machine Intell., vol. 11, pp. 567-585, June 1989.
[10] Ciarlet, Philippe G, Mathematical elasticity. North-Holland, 1988.
[11] P. C. Chou and N. J. Pagano, Elasticity: Tensor, Dyadic, and Engineering Approaches. Princeton, NJ: Van Nostrand, 1967.
[12] G. A. Nielson, "Scattered data modeling," IEEE Comput. Graphics Appl., vol. 13
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top