In this paper, we revisit the noise-reduction problem in the time domain and present a way to decompose the ltered speech into two uncorrelated (orthogonal) components: the desire...
Jingdong Chen, Jacob Benesty, Yiteng Huang, Tomas ...
Quantization-based watermarking schemes are vulnerable to amplitude scaling. Therefore the scaling factor has to be accounted for either at the encoder, or at the decoder, prior t...
Ivo D. Shterev, Reginald L. Lagendijk, Richard Heu...
Abstract. We deal with a map-labeling problem, named LOFL (Leftpart Ordered Flexible Labeling), to label a set of points in a plane with polygonal obstacles. The label for each poi...
In Continuous Speech Recognition (CSR) systems, acoustic and Language Models (LM) must be integrated. To get optimum CSR performances, it is well-known that heuristic factors must ...
Many watermarking algorithms embed the watermark into the image as contiguous non-overlapping tiles. This tiling structure forms an implicit synchronization template that can be r...
This paper deals with the adaptive variance scaling issue in continuous Estimation of Distribution Algorithms. A phenomenon is discovered that current adaptive variance scaling me...