Robustness of a multiscale scheme of feature points detection

We present a new scheme for feature points detection on a grey level image. I:ts principle is the study of the gradient phase signal along object edges and the characterization of the behavior across scales of the wavelet coefficients of this signal. The features points are determined as transition points of this signal. In the second part, we study the robustness of the detection scheme against changes of the acquisition parameters: the viewpoint and the zoom of the camera, the object rotation, the luminescence variation and noise. The results show the method efficiency: most of the points are still detected even if these parameters vary. (C) 2000 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.