Parameter-Transfer Learning for Low-Resource Individualization of Head-Related Transfer Functions

Abstract

Individualized head-related transfer functions (HRTFs) play an important role in accurate localization perception. However, it is a great challenge to efficiently measure continuous HRTFs for each person in full space. In this paper, we propose a parameter-transfer learning method termed PTL to obtain individualized HRTFs based on a small set of HRTF measurements. The key idea behind PTL is to transfer a HRTF generation model from other database to a target individual. To this end, PTL first pretrains a deep neural network (DNN)based universal model on a large database of HRTFs with the assist of domain knowledge. Domain knowledge is used to generate the input features derived from the solution to sound wave propagation equation at the physical level, and to design the loss function based on the knowledge of objective evaluation criterion. Then, the universal model is transferred to a target individual by adapting the parameters of a hidden layer of DNN with a small set of HRTF measurements. The adaptation layer is determined by experimental verification. We also conduct the objective and subjective experiments, and the results show that the proposed method outperforms the state-of-the-arts methods in terms of LSD and localization accuracy.

Pre-Trained-Based Individualization Model for Real-Time Spatial Audio Rendering System

“…Then, we compare the performance of our method with random selection, the methods in [16], [14] and [22] using 5 subjects in the test database. The random selection method randomly chooses a subject's HRTFs from the database as a target subject's HRTFs, where the database is the same as the one used in our model.…”

“…The individual weights outside the database are estimated from a small set of measurements, and then the HRTFs with highly directional resolution are obtained by combining the spatial basis functions and the weights. [14] proposes a parameter-transfer learning method to obtain individualized HRTFs based on a small set of HRTF measurements. However, this requires a priori measured HRTFs, which do not always exist for each subject.…”