This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: Abstract We propose a novel technique based on compressive sensing for expression-invariant face recognition. We view the different images of the same subject as an ensemble of intercorrelated signals and assume that changes due to variation in expressions are sparse with respect to the whole image. We exploit this sparsity using distributed compressive sensing theory, which enables us to grossly represent the training images of a given subject by only two feature images: one that captures the holistic (common) features of the face, and the other that captures the different expressions in all training samples. We show that a new test image of a subject can be fairly well approximated using only the two feature images from the same subject. Hence we can drastically reduce the storage space and operational dimensionality by keeping only these two feature images or their random measurements. Based on this, we design an efficient expression-invariant classifier. Furthermore, we show that substantially low dimensional versions of the training features, such as (i) ones extracted from critically-downsampled training images, or (ii) low-dimensional random projection of original feature images, still have sufficient information for good classification. Extensive experiments with publically-available databases show that, on average, our approach performs better than the state-of-the-art despite using only such super-compact feature representation. 1. Introduction Face recognition (FR) has been a highly active research area for many years. A typical approach involves two tasks: feature extraction and classification. Commonly- used feature extraction methods include subspace techniques such as principle component analysis (PCA or eigenface), independent component analysis (ICA), linear discriminant analysis (LDA or fisherface) and so on [1, 2] . With features extracted, classifiers based on techniques such as nearest neighbor and support vector machines can then be used to perform recognition. The above feature extraction methods are well-understood and in a sense have reached their maturity. Researchers are now looking for different methods and theories to address the persisting challenges in FR like expression, illumination and pose variation, and dimensionality reduction, etc. Reducing the space complexity and in particular the operational dimensionality of the classifier is vital for practical applications involving large databases. The recently-emerged Compressive Sensing (CS) theory [6,10,12-16] , while originally intended to address signal sensing and coding problems, has shown tremendous potential for other problems like pattern representation and recognition [3,4] , often beating the conventional techniques. In this paper we propose a new technique for face feature extraction and classification, based on the CS theory. We focus on addressing expression variation in FR. Expression-invariant FR is a challenging task owing to complex and varied nature of...
View Full Document
This note was uploaded on 05/28/2010 for the course EE EE564 taught by Professor Runyiyu during the Spring '10 term at Eastern Mediterranean University.
- Spring '10