Confinement of the infrared divergence for the Mumford process (Q861333)

The Mumford process is a stochastic distribution modulo constant and cannot be defined as a stochastic distribution invariant in law by dilations. The Mumford process has been introduced by \textit{D. Mumford} and \textit{B. Gidas} [Q. Appl. Math. 59, No.~1, 85--111 (2001; Zbl 1159.68598)] as the simplest process which can be generate images. In this paper, the author presents two explicit solutions. The first one consists in writing, in the frequency domain, as the sum of a radial term and an anti-radial term. Expanding the terms on a suitable orthonormal basis, the infrared divergence is carried by the radial term. The second solution is based on the construction of an adapted basis, the wavelet basis with pseudo-constant. It allows us to confine the infrared divergence on a smaller term than with the previous solution but the terms are now correlated. Note that there exists an orthonormal basis which provides us a confinement of the same order than the one given by the wavelet basis with pseudo-constant, but with decorrelated terms. But this `ideal' solution is not constructive [\textit{B. Vedel}, ``Règlement de la divergence infra-rouge dans des bases d'ondelettes adaptées'', Thèse de doctorat, Université de Picardie Jules Verne (2004)].