AvID is a multi-modal emotional speech database recorded simultaneously with different microphones and video camera. Our goal was to obtain a database to enable both the experiments in speaker identification/verification and detection of the emotional stage of persons involved in communication. Especially we pay attention to gather data involving spontaneous emotions and therefore obtain more realistic training and testing conditions for experiments. Spontaneous emotions were induced with specially planned scenarios including playing computer games and adaptive intelligence tests. Database consists from 12,5 hours of recordings from 17 speakers.
F.15 Development of a new information system/databases
COBISS.SI-ID: 7037012In 2009 the University of Surrey (UK) organized an international face recognition competition as part of the international IEEE/IAPR conference on biometrics (ICB'09). The goal of the competition was to assess the performance of stateoftheart face recognition systems and to identify future research direction in the field of facebased biometrics. Several established international research institutions participated in the competition, including Switzerland’s IDIAP, UK’s University of Surrey, Spain’s University of Vigo and University of Valencia, Amsterdam’s CWI, Britain’s University of Manchester and last but not least Slovenia’s University of Ljubljana. The competition was conducted on the challenging BANCA database using two different experimental configurations. On the more difficult of the two configurations our system (i.e., the system of the University of Ljubljana) resulted in the best performance and also achieved the overall best recognition results of the competition. When implementing the prototype face recognition system for the competition we gathered novel knowledge and important insights into the problems and challenges met when developing fully functional (deployable) face recognition systems.
B.03 Paper at an international scientific conference
COBISS.SI-ID: 7132756With our two systems we achieved 2nd place in the Intoxication and the Sleepiness SubChallengein the Interspeech 2011 Speaker State Challenge between 18 groups participating in the evaluation. Both systems are based on a Universal Background Model (UBM) in a form of a Hidden Markov Model (HMM), and the Maximum A Posteriori (MAP) adaptation. With the combination of our HMMUBMMAP derived supervectors and selected statistical functionals from the baseline feature set, we were able to surpass the baseline system in both subchallenges. By employing majority voting fusion of best systems we were able to further improve the performance. In the Intoxication SubChallengeour best result on the test set is 67.46%, and in the Sleepiness SubChallenge 71.28%.
B.03 Paper at an international scientific conference
COBISS.SI-ID: 8580692A novel 3D triangulation-based imaging method and device is proposed, which has the following the following advantages over known state-of-the-art methods: low energy consumption, enables uninterrupted usage of several devices in the same room, is robust to illumination changes and disturbancies. This patent was recognised as the most important achievement on the 2.15 Metrology field among the 2011 ARRS programs and projects.
F.32 International patent
COBISS.SI-ID: 8985172This paper describes current research and applications on speech technologies for the Slovenian language, as well as the complexity and interdisciplinary nature of the problem. Importance of Slovenian language for our national identity and problems and achievements concerning implementation of speech technologies systems for Slovenian language is emphasized.
F.30 Professional assessment of the situation
COBISS.SI-ID: 8160852