A fundamental problem in neural network research, as well as in many other disciplines, is finding a suitable representation of multivariate data, i.e. random vectors. For reasons...
Abstract. This paper presents a study of 25 structural features extracted from samples of grapheme `th' that correspond to features commonly used by forensic document examiner...
—Feature Extraction algorithm is a very important component of any retrieval scheme. We propose M-band Wavelet Transform based feature extraction algorithm in this paper. The MxM...
It is difficult to identify sentence importance from a single point of view. In this paper, we propose a learning-based approach to combine various sentence features. They are cat...
Speech recognition techniques have been developed dramatically in recent years. Nevertheless, errors caused by environmental noise are still a serious problem in recognition. Empl...