基于统计处理图像放大方法在人脸识别中应用

Application of an Image Magnification Method Based on Statistical Processing in Face Recognition

图像插值是数字图像处理中最基本、最重要的技术之一。文中设计提出一种新的边缘方向算法得到高分辨率图像的插值,并且把这个放大算法运用在人脸识别中。在很多视频监控中,尤其是当目标人脸离摄像头距离非常远,获得的目标人脸图像通常比较小,以至于难以对目标人脸图像进行正确的识别。文中首先提出了一种基于统计处理图像放大方法,使图像放大后更为清晰和易辨。然后应用主成分分析(PCA)和径向基函数网络(RBF)方法对放大后的人脸图像进行识别。经过多次实验,结果表明该识别方法在低分辨率人脸图像上有较好的识别效果,为人脸的实时识别提供了一种新途径。

Image interpolation is one of the most important technologies in digital image processing. It proposes an edge-directed interpolation algorithm for natural images to adapt the interpolation at a higher resolution and then use this method in face recognition. In many video surveillance systems, especially, when camera is very far away from the target object, the face image is usually small, which makes it difficult to recognize the human face correctly. It presents an image magnification method based on statistical processing, which makes the original image clearer and easier to distinguish. Then the principle component analysis （ PLA ） and the radial basis function neural network （RBF） are applied to human face recognition. The experimental results show that the new recognition method has a high recognition rate in low resolution face images. It provides a new way for real-time face recognition.