基于Gabor多方向特征融合与分块直方图的人脸表情识别方法

针对传统的Gabor特征表征全局特征能力弱以及特征数据存在冗余性的缺点, 提出一种新颖的采用Gabor多方向特征融合与分块直方图统计相结合的方法来提取表情特征. 为了提取局部方向信息并降低特征维数, 首先采用Gabor滤波器提取人脸表情图像的多尺度和多方向特征, 然后按照两个融合规则将相同尺度不同方向的特征融合到一起. 为了能够有效地表征图像全局特征, 将融合图像进一步划分为若干矩形不重叠且大小相等的子块, 分别计算每个子块区域内融合特征的直方图分布, 将其联合起来实现图像表征. 实验结果表明, 这种方法无论在计算量上还是识别性能上都比传统的Gabor滤波器组更具有优势. 该方法的创新处在于提出了两个Gabor多方向特征融合规则, 应用在JAFFE表情库上最高平均识别率达到98.24%, 表明其适用于人脸表情图像的分析.

Facial Expression Recognition Method Based on Gabor Multi-orientationFeatures Fusion and Block Histogram

In this paper, the Gabor multi-orientation fused features are combined with block histogram to extract facial expressional features in order to overcome the disadvantage of traditional Gabor filter bank, whose high-dimensional Gabor features are redundant and the global features representation capacity is poor. First, to extract the multi-orientation information and reduce the dimension of the features, two fusion rules are proposed to fuse the original Gabor features of the same scale into a single feature. Second, to represent the global features effectively, the fused image is divided into several nonoverlapping rectangular units, and the histogram of each unit is computed and combined as facial expression features. Experimental results show that the method is effective for both dimension reduction and recognition performance. The novelty of the method is to use two fusion rules to fuse multi-orientation Gabor features. The best average recognition rate of 98.24% is achieved in JAFFE database, which indicates this method is suitable for facial expression linebreak analysis.