1: \begin{abstract}

2: Most signal processing problems involve the challenging task of

3: multidimensional probability density function (PDF) estimation.

4: In this work, we propose a solution to this problem by using a family of Rotation-based

5: Iterative Gaussianization (RBIG) transforms.

6: The general framework consists of the sequential application of a

7: univariate marginal Gaussianization transform followed

8: by an orthonormal transform.

9: The proposed procedure looks for differentiable transforms to a known PDF so

10: that the unknown PDF can be estimated at any point of the original domain.

11: In particular, we aim at a zero mean unit covariance Gaussian for convenience.

12: 

13: 

14: RBIG is formally similar to classical iterative Projection Pursuit (PP) algorithms.

15: However, we show that, unlike in PP methods, the particular

16: class of rotations used has no special qualitative relevance in this context, since looking

17: for {\em interestingness} is not a critical issue for PDF estimation.

18: The key difference is that our approach focuses on the univariate part (marginal Gaussianization) of the problem rather than on the multivariate part (rotation).

19: This difference implies that one may select the most

20: convenient rotation suited to each practical application.

21: 

22: The differentiability, invertibility and convergence of RBIG are theoretically

23: and experimentally analyzed. Relation to other methods, such as Radial Gaussianization (RG), one-class support

24: vector domain description (SVDD), and deep neural networks (DNN) is also pointed out.

25: The practical performance of RBIG is successfully illustrated in a number of multidimensional

26: problems such as image synthesis, classification, denoising, and multi-information

27: estimation.

28: \end{abstract}

29: