有姿态变化的人脸图像识别方法论文.docVIP

编 号： 审定成绩： 重庆邮电大学 毕业设计（论文） 设计（论文）题目： 学院： 专 业 ： 班 级 ： 学 号 ： 指 导 教 师 ： 答辩组 负责人 ： 填表时间：年月 重庆邮电大学教务处制 人脸识别是一种研究以生物特征为基础的使用计算机实现的识别技术。在生物特征识别领域，人脸识别技术操作简单快速，结果直观，准确可靠，不需要人为的配合，正是这些优点使其成为一个热门学科。人脸识别在人工智能领域中占有极其重要的地位，是由于它自身的优越性以及在计算机视觉、图像处理、模式识别、多媒体、心理学等多个领域的广泛运用。 本文研究的重点正是在姿态变化下的多姿态人脸识别，研究测试样本出现姿态变化时的人脸识别的问题。 本文按照完整人脸识别流程来分析基于PCA(主成分分析)的人脸识别算法实现的性能,选用了ORL从人脸识别的实效性和可操作性两个角度，本文基于二维图像的正弦变换，其特点是可以实现快速的姿态校KNN（K近邻）分类器来进行人脸判别分类。最后比较有无姿态校正模块的人脸识别系统的准确率。 【关键词】 姿态正 PCA算法 近邻分类器 ABSTRACT Face recognition is a computer recognition technology, which is based on biological characteristics. In the field of biometric identification, the advantages of face recognition are fast and simple and the results are intuitive, accurate and reliable. It has become a popular subject for it does not need co-ordination. Due to its advantages, face recognition has been widely used in computer vision, image processing, pattern recognition, multimedia processing, psychology and other fields. So it plays an important role in the field of artificial intelligence. This study is to focus on pose-varied face recognition with the change of posture and research the face recognition problem when the postures of testing samples have changed. In practice, we merely get a minimal amount of training samples. If test samples have encountered the posture changing, the recognition rate of the traditional methods would reduce dramatically, which causes the system to be out of work. This paper focuses specifically on developing face recognition strategy of multi-gesture samples, which makes data easy to handle, and promoting the practical application of face recognition algorithms and programs. This paper analyzes the performance of face recognition algorithm which based on the PCA (Principal Component Analysis) and the experiments are conducted on the ORL face database. Considering the e