REGULARIZED MULTICHANNEL RESTORATION USING CROSS-VALIDATION

Multichannel images are the multiple image planes (channels) obtained by imaging the same scene using multiple sensors, The validity of mult ichannel restoration where both within- and between-channel relations are incorporated has already been established using both stochastic an d deterministic restoration filters, However, it has been demonstrated that stochastic multichannel filters are extremely sensitive to the e stimates of the between-channel statistics, In this paper we avoid the problems associated with multichannel stochastic filters by proposing deterministic multichannel filters that do not require any prior know ledge about either the statistics of the multichannel image or the noi se, Regularization based on the multichannel cross-validation function is used to obtain these filters, We examine their relation to multich annel linear minimum mean square error restoration filters and we prop ose a technique to estimate the variance of the noise, Finally, we sho w experiments where we test the proposed filters and noise variance es timator using color images. (C) 1995 Academic Press, Inc.