In machine learning, a deep belief network is a generative graphical model, or alternatively a class of deep neural network, composed of multiple layers of latent variables, with connections between the layers but not between units within each layer.
Researchers from UC San Diego and an external collaborator have developed a machine learning algorithm for interpolation and extrapolation of head related transfer function (HRTF) databases. This invention provides for neural networks that can interpolate the HRTF at all locations around the head with reduced spectral distortion. The technology holds promise for predicting individualized HRTFs.
Any application which uses HRTF, such as augmented/virtual reality, games, music distribution, and music performances can benefit from this algorithm.
This technology is patent pending and available for licensing and/or research sponsorship
HEAD RELATED IMPULSE RESPONSE INTERPOLATION; HRTF Interpolation, Deep Belief Networks, Binaural Audio