Research

Interests

  • Digital Signal Processing (in general).
  • Audio and Speech processing / enhancement.
  • Blind Source Separation.
  • Quality evaluation of Speech and audio.
  • Automatic Speech Recognition.
  • Robustness in Speech Recognition.
  • Biomedical Signal Processing.
  • Machine learning.
  • Computational Intelligence.
  • Telemedicine.
  • Communication systems.

Present research lines

Blind source separation

This technique aims at obtaining estimation of the sources that produced some measurable field, given a set of distributed measurements of the field itself. Specifically applied to audio sources, the field is a sound field measured trough several microphones located remotelly in some specified locations. The problem is known in the field as the "cocktail party problem", as in a party usually there are a lot of sound sources and yet human beings are able to focus their attention and listen to some specific source, like the person to wich one is having a conversation. This problem is very difficult to solve, as in the general case, the sound field is affected by the room characteristics (the reverberation phenomenon), the sources and sensors can be arbitrarily located, and in the most general case, the sources and/or sensors can be moving in the space. This is the subject of research in my PH D. thesis. I am exploring the use of Frequency Domain ICA to achieve the separation, overcoming the problems reported for this technique.

Statistical models for wavelet transforms

When using some domain transform to analyze signals in probabilistic models, it is usual to assume gaussianity of the resulting coefficients. However, when the transform used is a wavelet transform, this assumption does not hold. We are working in statistical models that represents better the properties of the wavelet transform, and their applications, with particular interests in denoising and speech recognition.

Robust Speech Recognition

The aim of this research is to obtain an automatic speech recognizer that is robust in front of different use conditions. In particular my research has been oriented toward giving the sistem robustness to noise, using speech enhancement techniques (like the blind source separation technique mentioned above), or using a set of characteristics that are intrinsically robust to noise (like the application of wavelets).

Present Research Projects

  • Member of Project "Digital Signal Processing: Statistical Models and Wavelet Applications". Funding: UNL. ID: CAI+D 2005 #012-72. Director: Dr. Diego Milone. National University of Litoral, Argentina.
  • Collaborator in Project "Non-Conventional Techniques Applied to Noise Reduction for Hearing Aids". Funding: ANPCyT-UNER. ID: PICT 2002 #12700. Director: Dr. María Eugenia Torres. National University of Entre Ríos, Argentina.
  • Student in project "Bioinformatics and Data Mining. Generative and Discriminative Models: Development and Applications". Funding: ANPCYT. ID: PAV2003-00127-00001. Director: Dr. Hermenegildo Cecatto. Agencia para la Promocion de la Ciencia y la Tecnologia. Argentina.

Past research lines

Telemedicine

I was interested in applications of telemedicine. My undergraduate thesis was devoted to the development of a system to track the location of ambulances using GPS, and simultaneously send biomedical signals (like ECG, pressure, etc.) to a central station to remotely monitor the patient status. This system worked with a VHF radio link. It was developed as a prototype, including a mobile device (to be mounted in the ambulance), a receiver device (to decode the data), and a computer software (to show the information and control the whole system).

Past research projects

  • Post Doctoral Fellow researcher in project "Speech Recognition in a Specified Space" belonging to "Knowledge Cluster Project", promoted by MEXT (Ministry of Education, Culture, Sports, Science and Technology). Director: Dr. Masuzo Yanagida. Doshisha University, Japan. 2003 and 2005.
  • Project “Development of a Voice Laboratory". Funding: UNER. ID: PID UNER #6062 (07/F062). Directors: Dr. Hugo Leonardo Rufiner and Dr. Diego Milone. 2002-2004.
  • Project "Automatic Speech Recognition". Funding: UNER. ID: UNER #6036 (07/F036). Director: Dr. Hugo Leonardo Rufiner. 1999-2002.
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-Share Alike 2.5 License.