Education

Ph.D. in artificial intelligence, acoustic signal processing (Feb. 2007-June 2010) at the Department of Electrical and Computer Engineering, Polytechnic School of the University of Patras under the supervision of Prof. Nikos Fakotakis, Prof. John Mourjopoulos and Prof. Evangelos Dermatas. Member of the AI Group of the Wire Communications Laboratory. I was accepted 1st/11 in the postgraduate program.

PhD Thesis

“Digital Processing and Automatic Classification of Environmental Sounds”, June 2010

Abstract: The dissertation introduces methods for automatic recognition of generalized sound events. More precisely, the dissertation proposes a series of novel methodologies for the following applications of the specific technology: a) recognition of urban environmental sound events, b) speech/music discrimination and c) acoustic surveillance of hazardous situations. As regards to the needs of the first application we propose the usage of principal component analysis for the post-processing of the acoustic parameters in combination with an hierarchical classification topology. With respect to the second application we introduce the concurrent usage of feature sets which belong to frequency and wavelet domains. These are fed to hidden Markov models for automatic analysis of radiophonic signals. As regards to the third application we propose a methodology for detection of atypical situations which is self-adaptive to different acoustic conditions. Additionally we investigate the novelty detection technology for acoustic surveillance of spaces using a database which was recorded under real-world conditions. Lastly the dissertation introduces the idea of temporal feature integration, a process which leads to high recognition rate. This methodology is a general scope one and can be adapted so as to fit the needs of various applications.

Diploma of Electrical and Computer Engineer (5 years), Department of Electrical and Computer Engineering, Polytechnic School of the University of Patras, Greece (2001-2006, graduated 7th/82).

Diploma Thesis

“Compression of multispectral images using visual perception criterions”, October 2006, rated as Excellent (grade 10/10). Low-level implementation of image compression algorithms and their comparison as regards to the case of multispectral images.