Blind Separation of Spatio-temporal Data Sources

Abstract

Abstract. ICA and similar techniques have been previously applied to either one-dimensional signals or still images. We consider the problem of blind separation of dynamic sources, i.e. functions of both time and two spatial variables. We extend the Sparse ICA (SPICA) approach and apply it to a sliding data cube, defined by the two dimensions of the visual scene and the extent in time over which the mixing problem can be considered to be stationary and linear. This framework and formalism are applied to two special problems encountered in two different fields: The first deals with separation of dynamic reflections from a desired moving visual scene, without having any a priori knowledge on the structure of the images and/or their statistics. The second problem concerns blind separation of 'neural cliques' from the background firing activity of a neural network. The approach is generic in that it is applicable to any linearly mixed dynamic sources.

Blind separation of a dynamic image source from superimposed reflections

“…In particular it has been successful in separating images from superimposed reflections [1], [2], [3], and in separation of tissues in MRI [12]. However, problems such as recovery of a landscape imaged through a layer of semi-transparent medium have not been previously addressed in this context and have hardly been discussed in the context of other image processing approaches and algorithms [5], [6].…”

“…In our previous study [1], we have shown simulation results for the separation of dynamic reflections. Here we further discuss this problem and show results of experimental data for separation of dynamic images.…”

“…still) images by means of techniques based on the Independent Component Analysis (ICA) approach [2], [3] and other methods based on the physics of the problem [4]. In our previous study [1], we have extended the Sparse ICA algorithm (SPICA) approach to dynamic images (i.e. sequences of images), by considering subsequences of data, that are to a good approximation stationary, as threedimensional data structures.…”

Natural Signal Classification by Neural Cliques and Phase-Locked Attractors

“…Following the concept of LSM, liquid-state was extended to spatio-temporal firing structure in a limited time-window, called neural clique [6], [7]. In the sequel, we implement neural cliques for mapping classes of signals to specific firing patterns.…”