Digital signal processing methods and apparatus

Abstract

This invention generally relates to methods, apparatus and computer program for processing digital data using non-negative matrix factorisation. A method of digitally processing data in a data array defining a target matrix (X) using non-negative matrix factorisation to determine a pair of matrices (F, G), a first matrix of said pair determining a set of features for representing said data, a second matrix of said pair determining weights of said features, such that a product of said first and second matrices approximates said target matrix, the method comprising: inputting said target matrix data (X); selecting a row of said one of said first and second matrices and a column of the other of said first and second matrices; determining a target contribution (R) of said selected row and column to said target matrix; determining, subject to a non- negativity constraint, updated values for said selected row and column from said target contribution; and repeating said selecting and determining for the other rows and columns of said first and second matrices until all said rows and columns have been updated.

Description

technical field [0001] The present invention generally relates to methods, apparatus, and computer programs for processing digital data using non-negative matrix factorization; embodiments of the invention are more particularly applicable to displays such as passive matrix OLED (organic light emitting diode) Multi-Line Addressing (MLA) technology. Background technique [0002] Techniques for eigenanalysis (decomposition of data into weighted groups of features) include vector quantization (VQ), principal component analysis (PCA), and nonnegative matrix factorization (NMF). Broadly speaking, the purpose of vector quantization is to enable the identification of prototypical features used to represent data by identifying prototypical features according to the group to which they most correspond. Principal component analysis seeks to construct sets of orthogonal basis functions that can then be combined to represent the data. Because vector quantization does not attempt to rep...

AI Technical Summary

Problems solved by technology

Principal component analysis and other techniques such as singular value decomposition (SVD) have the disadvantage that the prototype or basis functions may be negative and thus lack any physical reality

So-called "diagonalized" NMF techniques are limited to diagonally offset solutions (row-driven schemes), which lead to faster convergence, but provide little benefit in the context of multi-line addressing schemes

Images