Color Image Identification and Restoration

Image identification involves estimating the properties of an imperfect imaging system from an observed image prior to the restoration process. In this paper we present a novel identification technique for multichannel image processing using the maximum likelihood estimation (ML) approach. The image is represented as an autoregressive (AR) model and blur is described as a continuous spatial domain model, to overcome some major limitations encountered in other ML methods. Cross-spectral and spatial components, which are inherent to multichannel imaging systems, are also incorporated in the model to improve the overall performance. Also, it is shown that blur extent can be optimally identified from noisy color images that are degraded by uniform linear motion or out-of-focus blurs. The novelty of this approach in identifying the blur of multichannel images is a major contribution in producing visually acceptable results which is significant for higher processing levels.