Presented papers written in English and published in the Conference proceedings will be submitted for posting to IEEE Xplore.

Event program

Tuesday, 5/31/2016 9:00 AM - 12:30 PM, Camelia 2, Grand hotel Adriatic, Opatija Invited Speech  V. Struc (Faculty of Electrical Engineering, Ljubljana, Slovenia) Face Alignment: Addressing Pose Variability in Face Recognition Systems  Accepted Papers  1. D. Sáez-Trigueros, H. Hertlein, L. Meng (University of Hertfordshire, Hatfield, United Kingdom), M. Hartnett (IDscan Biometrics Ltd, London, United Kingdom) Shape and Texture Combined Face Recognition for Detection of Forged ID Documents  This paper proposes a face recognition system that can be used to effectively match a face image scanned from an identity (ID) doc-ument against the face image stored in the biometric chip of such a document. The purpose of this specific face recognition algo-rithm is to aid the automatic detection of forged ID documents where the photography printed on the document’s surface has been altered or replaced. The proposed algorithm uses a novel combination of texture and shape features together with sub-space representation techniques. In addition, the robustness of the proposed algorithm when dealing with more general face recognition tasks has been proven with the Good, the Bad & the Ugly (GBU) dataset, one of the most challenging datasets contain-ing frontal faces. The proposed algorithm has been complement-ed with a novel method that adopts two operating points to en-hance the reliability of the algorithm’s final verification decision. 2. L. Meluch, I. Tokárová (Slovak University of Technology, Bratislava, Slovakia), P. Farkaš (Slovak University of Technology and Pan European University, Bratislava, Slovakia), F. Schindler ( Pan European University, Bratislava, Slovakia) Simple Method Based on Complexity for Authorship Detection of Text  Authorship-identification is an actual research area with different applications. Particularly it is a question if only highly specialised and well equipped entities could identify an author or it could be made by broader community via exploiting text. In order to shed light on this question a very simple and computationally not demanding method for author identification was proposed based on lossless compression algorithms. In this paper another such method based on complexity is proposed and its verification on real data is presented. Results of experiments proved, that also such a simple method could be used successful identification of authors. 3. Z. Sun, L. Meng, A. Ariyaeeinia (University of Hertfordshire, Hatfield, United Kingdom), X. Duan, Z. Tan (Aalborg University, Aalborg , Denmark) Privacy Protection Performance of De-identified Face Images with and without Background  This paper presents an approach to blending a de-identified face region with its original background, for the purpose of completing the process of face de-identification. The re-identification risk of the de-identified FERET face images has been evaluated for the k-Diff-furthest face de-identification method, using several face recognition benchmark methods including PCA, LBP, HOG and LPQ. The experimental results show that the k-Diff-furthest face de-identification delivers high privacy protection within the face region while blending the de-identified face region with its original background may significantly increases the re-identification risk, indicating that de-identification must also be applied to image areas beyond the face region. 4. I. Filković, Z. Kalafatić, T. Hrkać (Fakultet elektrotehnike i računarstva, Sveučilište u Zagrebu, Zagreb, Croatia) Deep Metric Learning for Person Re-Identification and De-Identification  Large amounts of visual data are gathered from various surveillance systems across different places and times, and have to be processed in order to infer the current state of the world. One of the common problems in surveillance scenarios is person re-identification, the task of associating a person across different cameras. On the other hand, these scenarios raise privacy concerns, which lead to the need for person de-identification, i.e. concealing person identity. This task is related to the re-identification in two aspects: (i) multiple appearances of the same person could be de-identified in similar manner; and (ii) if we discover the features useful for re-identification, we could try to hide the identity by modifying those features. Re-identification can be addressed as a classification problem. The state-of-the-art classification methods are based on deep learning. In this paper we explore the applicability of the recently proposed Triplet network architecture to the person re-identification problem, by applying it on VIPeR dataset. We show that the network is able to learn useful feature-space embeddings, and analyze its benefits and limitations. 5. D. Marčetić, S. Ribarić (FER, Zagreb, Croatia) Deformable Part-based Robust Face Detection under Occlusion by Using Face Decomposition into Face Components  In this paper, we propose modifications of deformable part-based models in order to increase the robustness of face detection under occlusion. The modifications are: i) the tree, representing the deformable part-based model of the frontal face, which is partitioned into 11 subtrees representing face components; ii) the weight of each face component which is obtained based on the results of psychological experiments; iii) the introduction of new scoring functions and thresholds; and iv) a new procedure for robust face detection based on the valuation of scoring functions and thresholds. The experiment was performed only for frontal face images, and thus this work is used only as a proof of concept. Based on the encouraging experimental results, we conclude that the proposed method is suitable for extension to detect faces with different poses under occlusions. 6. P. Grd, M. Bača (Fakultet organizacije i informatike, Varaždin, Croatia) Creating a Face Database for Age Estimation and Classification  Age estimation is an important task in classifying face images. Human age classification and estimation can be defined in many ways, but this paper concentrates on age estimation and classification based on two-dimensional images of people's faces. The aging process affects the structure and appearance of face in many ways. The changes that occur are related to facial morphology and changes in the face texture. Some characteristics of facial morphology appear only in people of a certain age and change during the aging process. In order to create an accurate algorithm for age classification, appropriate datasets for training and testing are required. There is a small number of datasets available for age estimation and classification based on images of human faces. In order to develop a more accurate algorithm, this paper describes the process of creating a new database with facial images useful for age estimation and classification. 7. R. Singh, B. Raj, D. Gencaga (Carnegie Mellon University, Pittsburgh, United States) Forensic Anthropometry from Voice: An Articulatory-Phonetic Approach  This paper addresses a problem that is of paramount importance in solving crimes wherein voice may be key evidence, or the only evidence: that of describing the perpetrator. The term Forensic anthropometry from voice refers to the deduction of the speaker’s physical dimensions from voice. There are multiple studies in the literature that approach this problem in different ways, many of which depend on the availability of sufficient volumes of speech for analysis. However, in the case of many voice-based crimes, the voice evidence available may be limited. In such cases it is especially advantageous to regard the recorded signal as comprising multiple pieces of evidence. In this paper, we show how this can be done. We explain why, for any anthropometric measurement from speech, it makes sense to consider the contributions of each articulatory-phonetic unit independently of others, and to aggregate the deductions from them only in the aftermath. This approach is based on the hypothesis that the relative evidence given by different compositional units of speech can be more indicative of the anthropometric factor being deduced, than the evidence derived from the aggregate voice signal. We explain the applicability of this approach through experiments on standard speech databases.

Croatian Society MIPRO, IEEE Croatia Section, University of Zagreb, University of Rijeka in cooperation with COST Action IC1206 „De-identification for Privacy Protection in Multimedia Content“ organize this Special Session. The Session seeks to bring together researchers, professionals, and practitioners to present and discuss recent developments and challenges in Biometrics, Forensics, and Privacy De-identification in Multimedia Content. 

Slobodan Ribarić (Croatia)

Jean-François Bonastre, France
Bojan Cukic, USA
Patrizio Campisi, Italy
Joe Cannataci, Malta
Paulo Lobato Correia, Portugal
Jana Dittmann, Germany
Andrzej Drygajlo, Switzerland
Ivo Ipšić, Croatia
Carmen Garcia Mateo, Spain
Lily Meng, UK
Emilo Mordini, Italy
Ioannis Pitas, Greece
Bhiksha Raj, USA
Daniel Ramos, Spain
Tieniu Tan, China
Zheng-Hua Tan, Denmark
Isabel Trancoso, Portugal 
Marcos Faundez Zanuy, Spain

Topics of interest include (but are not limited to):

• De-identification methods for biometric identifiers in multimedia documents
• De-identification methods for soft- and non-biometric identifiers
• Applications and added value of de-identified data
• Reversible de-identification
• Ethical, societal and legal aspects of de-identification and reversible de-identification
• Face, fingerprint, hand geometry, palmprint and ear biometrics
• Behavioural biometrics
• Soft-biometrics
• Multi-biometrics
• Novel biometrics
• Biometric systems and applications
• Biometric evidence for forensic evaluation and investigation
• Audiovisual biometrics for forensics examination
• Soft biometrics for forensics examination
• Forensic behavioural biometrics
• Biometric analysis of crime scene traces and their forensic interpretation
• Ethical and societal implications of emerging forensics biometrics

Official language is English.

Slobodan Ribaric
University of Zagreb
Faculty of Electrical Engineering and Computing
Unska 3
HR-10000 Zagreb, Croatia

Phone: +385 1 612 99 52
Fax: +385 1 612 96 53
E-mail: slobodan.ribaric@fer.hr

Opatija, with its 170 years long tourist tradition, is the leading seaside resort of the Eastern Adriatic and one of the most famous tourist destinations on the Mediterranean. With its aristocratic architecture and style Opatija has been attracting renowned artists, politicians, kings, scientists, sportsmen as well as business people, bankers, managers for more than 170 years.

The tourist offering of Opatija includes a vast number of hotels, excellent restaurants, entertainment venues, art festivals, superb modern and classical music concerts, beaches and swimming pools and is able to provide the perfect response to all demands.

Opatija, the Queen of the Adriatic, is also one of the most prominent congress cities on the Mediterranean, particularly important for its international ICT conventions MIPRO that have been held in Opatija since 1979 gathering more than a thousand participants from more than forty countries. These conventions promote Opatija as the most desirable technological, business, educational and scientific center in Southeast Europe and the European Union in general.

For more details please look at www.opatija.hr/ and www.opatija-tourism.hr/.