EUSIPCO 2014

We just received the great news that 6 papers related to the project was accepted for the European Signal Processing Conference (EUSIPCO) in Lisbon, Portugal later this year (September).

The titles of the papers are as follows:

  • Near-field Localization of Audio: A Maximum Likelihood Approach
  • DOA and Pitch Estimation of Audio Sources Using IAA-based Filtering
  • A Broadband Beamformer Using Controllable Constraints and Minimum Variance
  • Robust Pitch Estimation Using an Optimal Filter on Frequency Estimates
  • Robust DOA Estimation of Harmonic Signals Using Constrained Filters on Phase Estimates
  • Spatio-Temporal Audio Enhancement Based on IAA Noise Covariance Matrix Estimates.

See you in Lisbon!

Visiting FAU

During December, Jesper Rindom Jensen was a visiting researcher at Friedrich-Alexander Universität (FAU), Erlangen-Nürnberg. At FAU, he visited Prof. Rudolf Rabenstein.

Collaboration on new topics within joint audio-visual research was started during the stay. More specifically, new ideas on audio-based localization in the near-field with unknown temperature/speed of sound was pursued, and are expected to be published during 2014. Audio-based localization is one of the two main ingredients in joint audio-visual localization methods.

Project start

In September 2013, Jesper Rindom Jensen was awarded a three year postdoc grant from the Danish Council for Independent Research | Technology and Production Sciences. The project started on October 1st, 2013 and will run until September 30th, 2016.

The research idea that will be pursued in this project, is to combine audio and visual information for localization of, e.g., speech sources. Knowing the location of speech sources is of utmost importance in numerous applications, but localization is a difficult task, when only based on audio information. This is due to reverberation, background noise, miscalibration, interferering sources, etc. As these phenomena do not appear in the same form in visual recordings obtained using a camera, a more robust speech location estimate can be obtained by fusing the audio and visual information. This approach will be considered in this research project.

More information is found in the “About” section.

Welcome!

Welcome to the website for the postdoc project “Localization and Tracking of Speech – A Joint Audio-Visual Approach”. The outcome of the research of this project will be highlighted on this webpage. More information about the project are found in the “About” section and in the following post.