Search

This paper introduces the wavelet transform, and based on which brings forward a new computer doable method. When an image is processed by multi-level 2-dimension dispersed wavelet transform, we can get two data information: the low frequency and the high frequency, after that, we add some random data to the high frequency, and then reconstruct the image. When testing, we use the same wavelet decomposition method to detect the correlation between the high frequency and the random data. This method can be used to avoiding the image forgery.

The research is conducted to develop a methodology to model the emotional contents of speech as a linear function of time and speech features. In this study, emotional type is coming from six basic emotions (anger, disgust, fear, joy, sadness and surprise). However, in the application is not defined as a simple word, is quantified using the dimensions valence and arousal, and the value of valance or arousal is expressed as some interval. Results demonstrate that the model is more excellent than others.