Voice And Image Processing, The Complexity Of The Analysis Method

At present, most of traditional methods of the signal processing are based on the linear theory. However, the signal is nonlinear. When the linear methods are used to process the nonlinear signal, the result is only utmostly approximate, and is not the essence of the signal. In this paper, we use the nonlinear method to process the nonlinear signal.The paper proposes the applications of complexity measure in the speaker verification and the automobile license plate location. From the laboratory result, the study gained a pleasing result.In the speaker verification, the study is developed in the paper:a) KC complexity character is presented in this paper, which reflects the essence of the speaker.b) Being based on the traditional MFCC (Mel Frequency Cepstrum Coeficient) and this new character (KC complexity character), designs the automatic speaker identification system. There are two kinds of speaker identification system (text-dependent and text-independent) in this paper.In the automobile license plate location, the study is developed in the paper:a) KC complexity character is presented in this paper, which reflects the character of the different modules. We distinguish the license plate region and the background region with KC complexity.b) Using KC complexity, we can distinguish the license plate region from the vertical stripe region, which can't be distinguished with frequency analysis.