1(1/ 2) -Dimension cepstrum feature analysis of pulse signals

Tremendous calculation is a big disadvantage of bicepstrum analysis of pulse signals; to solve this problem, this paper proposes an 1(1/ 2)-dimension cepstrum feature analysis method for recognition and classification of drug users´ pulse signals. This method analyzes the signal elements of human pulse signals, taking the three feature parameters: zero component S_x(0,0), value of H_s(p) and CSER as the feature components of pulse signals, which show obvious critical values between drug users and non-users. This method has been applied in the classification and recognition experiment on 15 users and 15 non-users and shows a high recognition rate of above 90%.