彩色图像的人脸检测与跟踪方法研究

【作者】 冯慧君；

【导师】 沈富可；

【作者基本信息】 华东师范大学 ， 计算机应用技术， 2003， 硕士

【摘要】 本文针对彩色图像的人脸检测跟踪问题，提出了一种比较新颖的方法。整个算法包括以下三个主要模块。 首先基于颜色和形状分割出人脸候选区域，其中，皮肤检测采用HSCC颜色空间方法，这种方法结合HSV，YCrCb两种颜色空间中的色度分量H，S，Cr，Cb，用模糊隶属度函数来决定像素点的可能程度，这种方法能有效地从彩色图像中检测出各种不同肤色的人脸，同时减少图像背景的影响。 接着通过从定位人脸相似区域中抽取眼睛，嘴巴，鼻子等脸部特征来证实人脸。对脸部特征抽取，我们应用脸部区域内的灰度信息。通过观察可知，眼睛和嘴巴等脸部特征相对其它的脸部区域较暗，我们可以通过积分投影和人脸分布特征初步定位眼睛、嘴巴、鼻子等的大致位置，然后根据眼睛特征，用方差投影精确定位人眼，最后再根据人脸分布特征依次定位人脸各器官。 在可靠地检测出人脸后，就可以进行人脸跟踪。跟踪过程通过主动轮廓来实现，轮廓用“蛇”(snake)来表示。蛇的外部能量由颜色信息决定，它推动蛇的元素snaxel内外移动。 本算法简单使用，实验表明了该算法的有效性和鲁棒性。更多还原

【Abstract】 In this paper we present a novel approach for segmentation and tracking of faces in color images. The whole algorithm is composed of three main modules.Segmentation of faces is performed by evaluating color and shape information. First, skin-like regions are determined based on Fuzzy HSCC by integrating the HS and CrCb spaces based on the concept of fuzzy membership function. Then regions with elliptical shape are selected as face hypotheses.The skin segmentation is robust against different lighting conditions and various types of skin color. They are verified by searching for facial features in their interior.Our approach for facial feature extraction is based on the observation that facial features differ from the rest of the face because of their low brightness. Thus, we detect facial features candidates by evaluating the lopographic greylevel relief of the face region. After obtaining an eye window, the eye can be located by using variance projection functions. Based on vertical symmetry distances between facial features and the assessment of each facial feature, we choose the best face constellation.After a face is reliably detected, it is tracked over time. Tracking is realized by using an active contour model. The exterior forces of the snake are defined based on color features. They push or pull snaxels perpendicular to the snake. Results demonstrate the efficiency and feasibility of this algorithm, which can be used in time critical system.更多还原