摘　　要
生物識(shí)別是最能體現(xiàn)計(jì)算機(jī)智能的特征之一，它在科研工作和日常生活應(yīng)用方面都引起了巨大的轟動(dòng)。在所有生物特征里面，人臉特征是最普遍和最容易獲取的，所以，人臉識(shí)別一直都是模式識(shí)別和圖像處理領(lǐng)域最熱門的課題之一。為了減弱光照、表情、姿態(tài)、攝像視角、外貌(例如頭發(fā)、眼睛)等變化對(duì)識(shí)別的影響，提高三維人臉識(shí)別準(zhǔn)確性，本文針對(duì)光照、姿態(tài)、表情等影響三維人臉識(shí)別的因素進(jìn)行了研究。使用基于球面諧波基圖像的光照補(bǔ)償算法減弱光照對(duì)人臉識(shí)別的影響，利用三維形變模型和部件技術(shù)相結(jié)合的人臉識(shí)別方法進(jìn)行多姿態(tài)人臉的識(shí)別，針對(duì)表情變化帶來的塑性變形采用基于泊松方程完成非中性人臉網(wǎng)格模型向中性模型的變形，提高帶表情模型與同類中性模型之間的相似度。這三種算法在提高三維人臉識(shí)別的性能上都有一定的幫助，對(duì)于三維人臉識(shí)別系統(tǒng)的研究有了一定的理論基礎(chǔ)。



Abstract
Biometrics is one of the best embodies the characteristics of computer intelligence, its research work and daily life applications have caused a great sensation. In all the biological characteristics of which, facial feature is the most common and most easily accessible, so, face recognition, pattern recognition and image processing has been the hottest topic in the field. In order to weaken illumination, facial expressions, gestures, camera angle, appearance (eg hair, eyes) and so the impact of change on the recognition and improve the accuracy of three-dimensional face recognition, this paper light, gesture, facial expression and other factors affecting the 3D face recognition study. Spherical harmonics basic image-based illumination compensation algorithm reduced the influence of light on face recognition using 3D deformable model and parts is combined with multi-pose face recognition face recognition, for the expression of changes in the plastic deformation Poisson equation based on the completion of non-neutral facial mesh model deformation to the neutral model, increased with the same neutral expression model and the similarity between models. The three dimensional face recognition algorithm to improve the performance of all have some help, for 3D face recognition system for a certain theoretical basis.



目　　錄

摘　　要 i
Abstract ii
第1章 緒　論 3
1.1本文工作的來源 3
1.2目的和意義 3
1.3國內(nèi)外進(jìn)展 4
1.3.1國外進(jìn)展 4
1.3.2國內(nèi)進(jìn)展 4
1.4本文工作的主要內(nèi)容 5
第2章 光照對(duì)三維人臉識(shí)別影響算法的研究 6
2.1基于球面諧波基圖像的人臉圖像光照估計(jì) 6
2.1.1球面諧波基圖像模型 6
2.1.2人臉的球面諧波基圖像近似 7
2.1.3光照估計(jì) 9
2.2人臉圖像光照補(bǔ)償 10
2.2.1紋理圖像 10
2.2.2差圖像 11
第3章 姿態(tài)對(duì)三維人臉識(shí)別影響的算法研究 13
3.1三維形變模型 13
3.2基于ASM的三維人臉分片 15
3.3基于三維部件技術(shù)的人臉識(shí)別 16
第4章 表情對(duì)三維人臉識(shí)別的影響的研究 20
4.1算法思想 21
4.2(Guidance-based constraint deformation) (GCD)模型 23
結(jié)　　論 26
參 考 文 獻(xiàn) 27
致　　謝 28
外文科技資料翻譯 29
英文原文 29
中文譯文 38