Krzysztof Kucharski

Face Recognition by Fisher and Scatter Linear Discriminant Analysis

Fisher linear discriminant analysis (FLDA) based on variance ratio is compared with scatter linear discriminant (SLDA) analysis based on determinant ratio. It is shown that each optimal FLDA data model is optimal SLDA data model but not opposite. The novel algorithm 2SS4LDA (two singular subspaces for LDA) is presented using two singular value decompositions applied directly to normalized multiclass input data matrix and normalized class means data matrix. It is controlled by two singular subspace dimension parameters q and r, respectively. It appears in face recognition experiments on the union of MPEG-7, Altkom, and Feret facial databases that 2SS4LDA reaches about 94% person identification rate and about 0.21 average normalized mean retrieval rank. The best face recognition performance measures are achieved for those combinations of q,r values for which the variance ratio is close to its maximum, too. None such correlation is observed for SLDA separation measure.

Image Object Localization by AdaBoost Classifier

AdaBoost as a methodology of aggregation of many weak classifiers into one strong classifier is used now in object detection in images. In particular it appears very efficient in face detection and eye localization. In order to improve the speed of the classifier we show a new scheme for the decision cost evaluation. The aggregation scheme reduces the number of weak classifiers and provides better performance in terms of false acceptance and false rejection ratios.

Cascade of Dual LDA opearators for face recognition

In this paper, we propose a cascade of Dual-LDA (DLDA) operators for Face Recognition. We show that such an approach results in efficient and low-dimensional feature space for face representation with enhanced discriminatory power. Comparative results to classical LDA and cascade of classical LDA algorithms are presented, showing significantly improved performance. A theoretical analysis for Fisher and DLDA is also presented. Experimental evaluation of the proposed FR algorithm, conducted on MPEG test set with over 8000 images of 929 individuals, shows state-of-the-art performance.

AUTOMATIC FACE SYNTHESIS AND ANALYSIS. A QUICK SURVEY

Considerable interest has been received in automatic face synthesis and analysis over the last three decades. This paper surveys the current state of the art in face synthesis, and also presents face detection and eye detection selected algorithms along with facial feature extraction approach based on using Harris corner detector in face analysis.

CASCADE OF OPERATORS FOR FACIAL IMAGE RECOGNITION AND INDEXING

A cascade of linear and nonlinear operators is designed for facial image indexing and recognition. We show that such an approach results in efficient and low- dimensional feature space for face representation with enhanced discriminatory power. Experimental evaluation of the proposed FR algorithm was conducted on MPEG test set with over 8000 images of about 1000 individuals.

Two step appearance-based approach for fast and reliable face localization

Among many face detection methods the appearance-based ones have proved to be the most accurate and in particular the AdaBoost cascade algorithm is both accurate and a very fast technique. The high speed of the detector is a crucial parameter in many face detection applications, e.g. the face recognition. In this paper the central two-step detector is presented which is a serial connection of the cascade of the extended weak classifiers and the AdaBoost cascade. The cascade of the extended weak classifiers is a novel concept that accelerates the detection speed to a high degree. The second introduced novelty is the verification of the detection results with another AdaBoost cascade to push a number of false acceptances down to the extremely low level.

Feature space reduction for face recognition with dual linear discriminant analysis

Linear Discriminant Analysis (LDA) is widely known feature extraction technique that aims at creating a feature set of enhanced discriminatory power. It was addressed by many researchers and proved to be especially successful approach in face recognition. The authors introduced a novel approach Dual LDA (DLDA) and proposed an efficient SVD-based implementation controlled by two parameters. In this paper DLDA is analyzed from the feature space reduction point of view and the role of the parameters is explained. The comparative experiments conducted on facial database consisting of nearly 2000 individuals show superiority of this approach over class of feature selection methods that choose the features one by one relying on classic statistical measures.

Dual LDA - an effective feature space reduction method for face recognition

Linear discriminant analysis (LDA) is a popular feature extraction technique that aims at creating a feature set of enhanced discriminatory power. The authors introduced a novel approach dual LDA (DLDA) and proposed an efficient SVD-based implementation. This paper focuses on feature space reduction aspect of DLDA achieved in course of proper choice of the parameters controlling the DLDA algorithm. The comparative experiments conducted on a collection of five facial databases consisting in total of more than 10000 photos show that DLDA outperforms by a great margin the methods reducing the feature space by means of feature subset selection.