Search

This paper proposes a single-channel image blind restoration using iterative principal components analysis (PCA) to improve the quality of restoration. Previously proposed PCA approaches for blind restoration have a lot of problems. For example, the process of boosting high-frequency components would be improvable, no numerical evaluation has been performed, and etc‥ Generating an ensemble by means of Gaussian filter application, discussed in this paper, could improve to extract the high frequency components which had been lost. Furthermore, iterative PCA boosts the high frequency components. Our proposed method is applied to a restoration example of atmospheric turbulence-degraded imagery, and we verified to improve restoration quality in comparisons with conventional methods. For demonstrating comparative experiments, simulations have been conducted. From the results, we can confirm that the proposed method gives higher PSNR as well as SSIM than the conventional methods. [ABSTRACT FROM PUBLISHER]

A neural network for two-dimensional visual pattern learning and classification is outlined. The new architecture combines the important aspects of two previously developed network designs, simultaneously taking advantage of the unique properties of both. The structure of the Neocognitron network is incorporated to allow shift-invariant and partial scale-invariant recognition, while the top-down attentional and matching mechanisms found in the adaptive resonance theory (ART) model are used to solve the stability-plasticity dilemma. The new network is self-organizing, shift invariant and able to switch automatically between its stable and plastic modes. Computer simulation results for a group of edge extracted patterns are detailed. The neural design uses viable neural mechanisms similar to those thought to exist in biological neural systems [ABSTRACT FROM PUBLISHER]