Coding time-varying signals using sparse, shift-invariant representations
Proceedings of the 1998 conference on Advances in neural information processing systems II
Efficient Coding of Time-Relative Structure Using Spikes
Neural Computation
Binaural sound localization based on sparse coding and SOM
IROS'09 Proceedings of the 2009 IEEE/RSJ international conference on Intelligent robots and systems
Hi-index | 0.00 |
This paper presents a binaural sound source localization method using both a sparse coding and a self-organizing map (SOM) in real-time system. We use the sparse coding for feature extraction to estimate the azimuth of sound source. It is used for decomposing input sound signals into three components such as time, frequency and magnitude. Therefore, although the frequency characteristic of ITD (Interaural Time Difference) is changed by shape of head, we utilized it to estimate the azimuth of the sound source considering the time-frequency features simultaneously. Then we adapted the SOM to estimate the azimuth of sound source which is a type of artificial neural networks. This system is constructed by open-source software, Flowdesigner, which gives us a data-flow oriented developmental environment for efficient real-time system.