Generalized algorithm of detection and classification in uterine electromyography signal
Abstract
In this paper, we present a method of detection and classification of events well adapted to uterine EMG processing. A multidimensional method of detection will be presented. It is based on the sequential computation of the likelihood ratio after signal decomposition on pertinent scales using wavelet transform. Hypothesis rejection is achieved using variance covariance matrices computed from the scales. This approach leads to different detection and isolation delays, the former being defined by the detection threshold, the latter depending on the estimation time of the covariance matrix. This method is adaptive and allows event detection without necessarily returning to the null hypothesis H/sub 0/. It has been applied on uterine EMG and gives satisfactory results in both detection and classification.