Abstract:We address the problem of blind separation of multiple source layers from their linear mixtures with unknown mixing coefficients and unknown layer motions. Such mixtures can occur when one takes photos through a transparent medium, like a window glass, and the camera or the medium moves between snapshots. To understand how to achieve correct separation, we study the statistics of natural images in the Labelme data set. We not only confirm the well-known sparsity of image gradients, but also discover new joint behavior patterns of image gradients. Based on these statistical properties, we develop a sparse blind separation algorithm to estimate both layer motions and linear mixing coefficients and then recover all layers. This method can handle general parameterized motions, including translations, scalings, rotations, and other transformations. In addition, the number of layers is automatically identified, and all layers can be recovered, even in the underdetermined case where mixtures are fewer than layers. The effectiveness of this technology is shown in experiments on both simulated and real superimposed images.

Natural Gradient Approach To Blind Separation Of Over- And Under-Complete Mixtures

Natural Gradient Algorithm for Blind Separation of Overdetermined Mixture with Additive Noise.

Geometrical Structures of FIR Manifold and Multichannel Blind Deconvolution

A New Approach to Blind Separation of Overdetermined Mixtures

Conjugate Gradient Approach to Blind Separation of Temporally Correlated Signals

Blind separation of convolutive mixtures based on second order and third order statistics

Geometrical structures of FIR manifold and their application to multichannel blind deconvolution

Blind Separation of Superimposed Moving Images Using Image Statistics

Research on Blind Source Separation Based on Natural Gradient Algorithm

Blind Source Separation in Post-nonlinear Mixtures Using Natural Gradient Descent and Particle Swarm Optimization Algorithm

Online Algorithm of Blind Source Separation Based on Conjugate Gradient Method

Approach of blind source separation based on experience character function and its application

Blindly separating mixtures of multiple layers with spatial shifts

Blind Separation and Extraction of Binary Sources.

Blind Separation of Filtered Sources Using State-Space Approach

Semiparametric Approach to Multichannel Blind Deconvolution of Nonminimum Phase Systems

Blind Separation of Superimposed Images with Unknown Motions

New Geometric ICA Approach for Blind Source Separation

Blind separation of locally smooth images based on genetic algorithm

A New Mixing Matrix Identification Algorithm for Underdetermined Blind Source Separation

New Blind Source Separation Method Based on Genetic Algorithm