An Improved Unsupervised Single-Channel Speech Separation Algorithm for Processing Speech Sensor Signals

As network supporting devices and sensors in the Internet of Things are leaping forward, countless real-world data will be generated for human intelligent applications. Speech sensor networks, an important part Things, have numerous application needs. Indeed, can further help applications to provide higher quality services, whereas this may involve considerable noise data. Accordingly, speech signal processing method should urgently implemented acquire low-noise effective Blind source separation enhancement technique refer one representative methods. However, unsupervised complex environment, only presence a single-channel signal, many technical challenges imposed on achieving multiperson mixed separation. For reason, study develops CNMF+JADE, i.e., hybrid combined with Convolutional Non-Negative Matrix Factorization Joint Approximative Diagonalization Eigenmatrix. Moreover, adaptive wavelet transform-based is proposed, capable adaptively effectively enhancing separated signal. The proposed aimed at yielding general efficient algorithm acquired by sensors. revealed from experimental results, TIMIT sources, extract target speaker tiny training sample. highly robust, technically most