Attention-based Convolutional Neural Network for Sound Source Localization (ACNN-SSL) Using Octagonal Microphone Array

Sound source localization in noisy environments using microphone arrays is a challenging task that has attracted the attention of many researchers. Using deep learning approaches by reformulating the sound source localization (SSL) obtained superior performance in these classification tasks. In this paper, a new model of a convolutional neural network with specific extracted features is introduced to infer the direction of arrival (DOA) of a sound source in noisy and reverberant conditions. A concatenation of time domain and frequency domain features is extracted from sounds. The results depicted that the proposed model has superior performance than the other systems from the literature by achieving the highest accuracy in DOA estimation.

کلیدواژه ها:

Convolutional Neural Network ، Deep Learning ، Octagonal Microphone Arrays ، Sound Source Localization