The PhD (Pretty helpful Development functions for) face recognition toolbox is a collection of Matlab functions and scripts intended to help researchers working in the field of face recognition. The toolbox includes implementations of some of the most popular face recognition techniques, such as Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Kernel Principal Component Analysis (KPCA), and Kernel Fisher Analysis (KFA). It features functions for Gabor filter construction, Gabor filtering, and all other tools necessary for building Gabor-based face recognition techniques. In addition to the listed techniques there are also a number of evaluation tools available in the toolbox, which make it easy to construct performance curves and performance metrics for the face recognition technique you are currently assessing. These tools allow you to compute ROC (Receiver Operating Characteristics) curves, EPC (Expected performance curves) curves, and CMC (cumulative match score curves) curves. Most importantly (especially for beginners in this field), the toolbox also contains several demo scripts that demonstrate how to build and evaluate a complete face recognition system. The demo scripts show how to align the face images, how to extract features from the aligned, cropped and normalized images, how to classify these features and finally how to evaluate the performance of the complete system and present the results in the form of performance curves and corresponding performance metrics. The toolbox, which is currently available from its on-line repositories on Matlab Central, the face recognition homepage (http://www.face-rec.org/) and the toolbox’s homepage (http://luks.fe.uni-lj.si/sl/osebje/vitomir/face_tools/PhDface/), was downloaded several thousand times since its introduction in 2012. The toolbox ships with a user manual, which is also listed in the Slovenian bibliographic database COBISS [COBISS.SI-ID 8967764].
F.23 Development of new system-wide, normative and programme solutions, and methods
COBISS.SI-ID: 8967508The alignment of the facial region with a predefined canonical form is one of the most crucial steps in a face recognition system. Most of the existing alignment techniques rely on the position of the eyes and, hence, require an efficient and reliable eye localization procedure. In the lecture we introduce a novel technique for this purpose, which exploits a new class of correlation filters called Principal directions of Synthetic Exact Filters (PSEFs). The proposed filters exhibit desirable properties, such as relatively short training times, computational simplicity, high localization rates and real time capabilities. We present the theory of PSEF filter construction, elaborate on their characteristics and finally develop an efficient procedure for facial landmark localization using several PSEF filters. The effectiveness of the developed technique is demonstrated on the task of eye localization using more than 40000 facial images pooled from the FERET and LWF databases. The results of our experiments suggest that the PSEF filters produce significantly better localization results than, for example, the Haar-cascade object detector, while ensuring a more than 10-fold improvement in the processing time. The lecture is based on the ERK paper »Advanced correlation filters for facial landmark localization« [COBISS.SI-ID 9381716].
B.04 Guest lecture
COBISS.SI-ID: 9402452In 2013 the Swiss research institut Idiap organized a face recogniton competition as part of the IAPR International Conference on Biometrics (ICB). The main goal of the competition was to identify the best techniques, methods and most promising research directions for face recognition systems deployable in uncontrolled environments. The competition attracted stong international competition from different parts of the world. Each participating institution was asked to provide recognition results of one recognition system based on which the ranking of the copetition was established. As the experimenal database, Idiap's MOBIO database was selected. The database was recorded in different environments using various mobile devices and, therefore, reflect conditions encounetered in real-world settings. We entered the competition with a recognition system developed in the scope of the post-doctoral project BAMBI and achieved the best recognition results and consequently the first place in the competition.
E.02 International awards
COBISS.SI-ID: 10176852Similarity scores, which form the basis for identity inference in biometric verification systems, typically exhibit statistical variations. These variations are caused by so-called miss-matched conditions, in which the enrollment and probe samples were acquired, and are common to most application domains of biometric verification systems ranging from forensics to smart-home environments. To mitigate these variations, score normalization techniques are usually used. Examples of these techniques include the z-norm, the t-norm or the zt-norm. In this paper we study two-step normalization techniques, such as the zt-norm, and propose a new way of implementing such techniques. Specifically, we propose to implement the first step of the two-step procedure off-line in a non-parametric manner, while the second step is kept unchanged and, hence, performed parametrically. As shown in our face verification experiments, the proposed composite scheme can improve upon the performance of parametric normalization techniques, without an increase in computational complexity, as this is the case with pure non-parametric normalization techniques.
B.03 Paper at an international scientific conference
COBISS.SI-ID: 9757524The paper introduces a novel approach to face recognition based on the recently proposed low-dimensional probabilistic linear discriminant analysis (LD-PLDA). The proposed approach is specifically designed for complex recognition tasks, where highly nonlinear face variations are typically encountered. Such data variations are commonly induced by changes in the external illumination conditions, viewpoint changes or expression variations and represent quite a challenge even for state-of-the-art techniques, such as LD-PLDA. To overcome this problem, we propose here a patch-wise form of the LD-PLDA technique (i.e., PLD-PLDA), which relies on local image patches rather than the entire image to make inferences about the identity of the input images. The basic idea here is to decompose the complex face recognition problem into simpler problems, for which the linear nature of the LD-PLDA technique may be better suited. By doing so, several similarity scores are derived from one facial image, which are combined at the final stage using a simple sum-rule fusion scheme to arrive at a single score that can be employed for identity inference. We evaluate the proposed technique on experiment 4 of the Face Recognition Grand Challenge (FRGCv2) database with highly promising results.
B.03 Paper at an international scientific conference
COBISS.SI-ID: 9890644