Artifact reduction based on mutual information measures
Ming-Te Wu · 2016
Reduction of artifact effect has become indispensable standard in image reconstruction for the production of multimedia device. Since the initial use of artifact reduction in image reconstruction, much effort has been put into further increasing its performance, while decreasing the benefit required. Unfortunately, such development efforts have shown diminishing return in recent years and further significant improvement cannot be expected. This promoted many researchers to seek alternative processes for the restore of artifact reduction. We explored the use of mutual information with expected log likelihood measure for the image restore. Our results showed that the proposed measures are a good way for image reconstruction that produces excellent performance. Furthermore, the use of expected log likelihood on image coding process produced higher efficiency. Fine adaptation of expected log likelihood provided efficiency that was superior to those performed by the conventional methods.