Robust spatial subtraction array with independent component analysis for speech enhancement
Yu Takahashi, Tomoya Takatani, Hiroshi Saruwatari, Kiyohiro Shikano · 2007
In this paper, we propose a new spatial subtraction array (SSA) structure which includes independent component analysis (ICA)-based noise estimator. Recently, SSA has been proposed to realize noise-robust hands-free speech recognition. In SSA, noise reduction is achieved by subtracting the estimated noise power spectrum from the noisy speech power spectrum. The conventional SSA uses null beamformer (NBF) as a noise estimator, but NBF suffers from the adverse effect of microphone-element errors and room reverberations in real environments. To improve the problem, we newly replace NBF with ICA which can adapt its own separation filters to the element error and the reverberation. The affections by the element error and the reverberation can be mitigated in the proposed ICA-based noise estimator. Experimental results reveal that the accuracy of noise estimation by ICA outperforms that of NBF, and speech recognition performance of the proposed method overtakes that of the conventional SSA.