A combined approach for object detection and deconvolution

The Multiscale Vision Model is a recent object detection method, based on the wavelet transform. It al- lows us to extract all objects contained in an image, what- ever their size or their shape. From each extracted object, information concerning flux or shape can easily be deter- mined. We show that such an approach can be combined with deconvolution, leading to the reconstruction of de- convolved objects. We discuss the advantages of this ap- proach, such as how we can perform deconvolution with a space-variant point spread function. We present a range of examples and applications, in the framework of the ISO, XMM and other projects, to illustrate the eectiveness of this approach.