Abstract: We present an analysis of a 2-D planar shape descriptor based on the Discrete Wavelet Transform (DWT), as an approach for compact object representation in multimedia applications. To generate the wavelet shape descriptor (WD), a partial 1-D DWT is applied on the vertical and horizontal components of a parameterized closed curve that represents the contour of an object of interest. With this shape representation tool and based on a multiresolution analysis, the curve can be reconstructed on a scale-by-scale basis to the desired degree of approximation. We show that contours of typical multimedia video objects are well represented with a small number of the largest magnitude wavelet coefficients, allowing for an efficient object shape representation. Some issues on the selection of the coefficients are studied. A comparison between the performance of the WD using different wavelet filters and a Fourier shape descriptor is presented, for synthetic and natural contours. The potential use of this descriptor in a multimedia application framework is discussed.