US 5,351,079 A · Color balance adjusting apparatus using a decorrelating neural network

Abstract

A color balance adjusting apparatus for adjusting color imbalance due to the difference in color temperature between a preferable standard illuminant and an illuminant under which a color image is obtained, thereby making the colors of the image substantially identical to those obtained under the preferable standard illuminant. This apparatus comprises a decorrelating neural network for receiving three color component signals correlating with one another and indicative of an image, and minimizing the correlation thereamong, the network having learned so as to minimize the correlation among signals indicative of an image obtained under an illuminant, and a converter for mapping the output of the decorrelating means, into a space of the input image signals, with the use of the inverse matrix of a transfer matrix of the neural network having learned under the preferable standard illuminant. The output of the converter is generated as a signal obtained after white balance adjusting.

Claims18

Dependent claims

A color balance adjusting apparatus comprising:

decorrelating means for receiving first color component signals which indicate a color image obtained under an illuminant and which are correlated with one another and converting the first color component signals into second color component signals of reduced correlation, the decorrelating means comprising a neural network which is formed of neuron units having feedback connections and being learned so as to reduce the correlation among the first color component signals;
and
conversion means for converting the second color component signal output from the decorrelating means into signals indicating a color image having the same color balance as a color image obtained under a standard illuminant, said conversion means utilizing an inverse matrix of a transfer matrix of the neural network having been learned under the standard illuminant.

A method of adjusting color balance by means of a decorrelating neural network for reducing the correlation among input signals and converting means, comprising the steps of:

supplying the decorrelating neural network with first color component signals which are correlated with one another and which indicate a color image obtained under a standard illuminant, and causing the decorrelating neural network to learn so as to reduce the correlation of the first color component signals;
storing an inverse matrix of a transfer matrix of the decorrelating neural network having been learned;
supplying the decorrelating neural network with second color component signals which are correlated with one another and which indicate a color image obtained under a given illuminant with a given color temperature, and causing the decorrelating neural network to learn so as to reduce the correlation of the second color component signals;
supplying the output of the decorrelating neural network supplied with the second color component signals and the stored inverse matrix with the converting means, and multiplying the output of the decorrelating neural network by the inverse matrix, with the result that the output of the converting means indicates an image having the same color balance as that of the color image obtained under a standard illuminant.

1 more element

Dependencies are inferred from the claim text.

Color balance adjusting apparatus using a decorrelating neural network

Life of this patent

Color balance adjusting apparatus using a decorrelating neural network

Life of this patent

Abstract

Claims18