期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2014
期号：ICETS
页码：638
出版社：S&S Publications
摘要：The objective of the project is totransform the samples in different modality to acommon feature space. The discriminant features forthe modalities are well aligned so that the comparisonbetween them is Possible. In this paper, weproposedmethod to recognize the heterogeneous facerecognition. The prototype subjects (i.e., the trainingset) have an image in each modality (probe andgallery), and the similarity of an image is measuredagainst the prototype images from the correspondingmodality. Initially filtered with three different imagefilters such as Difference of Gaussians, CSDN (CenterSurround Devise Normalization), and Gaussian filter.After this process we have to identify the SIFT, CLBPfeatures. It is an algorithm in computer vision todetect and describe local features in images. Thesimilarity between the two pattern images willidentify by the kernel similarity. Finally the identifiedimage will be retrieved from the database.
关键词：Discriminant analysis; Feature;descriptors; Heterogeneous face recognition; Kernel;similarityprototypes; and View sketch
