Sieves and wavelets: multiscale transforms for pattern recognition

In this paper a scalelposition decomposition that is an alternative to wavelets is described. The nonlinear decomposition, called the datasieve, is appropriate for isolating and locating the position of objects with sharp edges arising from nonlinear events such as occlusion. It can represent structural information in a way that is independent of spatial frequency, has different uncertainty tradeoffs, and can be used for scale, position and contrast independent pattern recognition.