Active Noise Cancellation using Deep learning

Malihe Hassani¹

1) گروه مهندسی برق و کامپیوتر، واحد بابل،دانشگاه آزاد اسلامی،بابل،ایران

IOI Code : ABGZ-ACABAC
Publication : The third international conference on new approaches in engineering and basic sciences(icnabs.ir)
Abstract :
    Active Noise Cancellation (ANC) is one of the most effective ways of reducing noise. The active noise reduction headphone is the most successful application of active noise control. Traditional active noise control methods are based on adaptive signal processing with the least mean square algorithm as the foundation. They are linear systems and do not perform satisfactorily in the presence of nonlinear distortions. In this project, ANC is formulated as a supervised learning problem and a deep learning approach, called deep ANC is proposed. Hybrid Active noise cancellation techniques which is the combination of feed forward and feedback techniques, in this project. Large scale multi conditioning is trained to achieve good generalization and robustness against a variety of noises. The goal of ANC systems is to generate an anti-noise with the same amplitude and opposite phase of the primary (unwanted) noise to cancel the primary noise. A Convolutional Recurrent Network (CRN) is trained to estimate the real and imaginary spectrograms of the cancelling signal from the reference signal so that the corresponding anti-noise can eliminate or attenuate the primary noise in the ANC system.

Keywords : deep anc convolutional recurrent network librosa denoise