ABSTRACT
CONTENTS 9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 9.2 Image and Video Observation Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300 9.3 Regularization of Ill-Posed Inverse Problems . . . . . . . . . . . . . . . . . . . . . . . 302 9.4 Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 9.5 Summary and Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
Image and video enhancement involves the extraction of high-frequency details from an observed visual data set. In this chapter, several open research problems are investigated, including region-of-interest image magnification from a digital still image, the postprocessing of still image compression artifacts, and the integration of multiple digital video frames/fields to generate a superresolved video still. Primary real-world applications include the enhancement of remote sensing, reconnaissance, and surveillance images and video. The overall goal of this research is to perform enhancement on the image data as accurately as possible, so computational efficiency is not an issue. Specifically, iterative nonlinear enhancement techniques will be considered for their edge-preserving properties, although linear algorithms related to least squares are generally much faster and easier to implement. However, linear estimation results in smooth (low-frequency) solutions, which do not contain sharp, discontinuous structures.
