语音识别说话人自适应研究现状及发展趋势

说话人自适应是提高非特定人语音识别系统识别性能的有效手段.本文介绍了说话人自适应研究的现状,包括自适应的不同方式和不同算法,并详细介绍了目前应用最为广泛的MLLR算法和MAP算法.本文还给出了对说话人自适应研究发展趋势的预测.

Evaluating the Visibility of Image Watermarking in the DCT Domain Based on Laplacian Model

Digital watermarking is a key technique for protecting intellectual property of digital media.As a number of methods have been proposed in recent years to embed watermarks in images for various applications,evaluation of watermarking algorithms becomes more and more important.Watermark visibility,which can be measured by signal to noise ratio (SNR) or peak signal to noise ratio (PSNR),is one of the major performance indexes of watermarking algorithms.In this paper,based on the Laplacian distribution model of AC DCT coefficients,we deduce theoretical relationship between the scaling parameter in some typical watermarking algorithms and the degradation of watermarked images.Experimental results show that the evaluated error of SNR and PSNR is less than 1 dB.