Curvelet / Surfacelet comparison

Curvelets and Surfacelets are two transforms that aim to achieve a multiscale and a multidirectional decomposition of arbitrary N-dimensional ($N>=2$) signals. While both transforms are Fourier-based, their construction is intrinsically different. In this talk we will give and overview of the construction of the two transforms, and explore their properties such as frequency domain / spatial domain coherence, sparsity, redundancy and computational complexity.