In this paper, a multiscale wavelet-based algorithm for matching stand-alone shapes is developed. The algorithm uses the Dyadic Wavelet Transform (DWT) to decompose a shape?s boundary into multi-scale levels. Features are extracted by calculating the curve moment invariants of the approximation coefficients. If the measured dissimilarity is small, then the shapes are globally similar. Local similarity is investigated by calculating the normalized cross correlation of the 1-D triangle area representation of the detail coefficients. The presented algorithm not only finds similar shapes, but it also can easily distinguish between seemingly similar shapes. The algorithm is invariant to the affine transformation and to the starting point variation of the shape contour.