Recently we published a new image compression method utilizing a combination of discrete cosine transform and least squares interpolation method. In this talk, we will present a discussion of the mathematical background, outline of the approach, complexity computations, pseudocode, and an explanation of how to implement the algorithm for applications that require the coded bits to be binary streams. We then provide the results, including comparisons to many recently published works. Also, we compare our work versus both JPEG and JPEG 2000, where we have significant improvement. The results indicate positive progress and effectiveness of the new approach in terms of comparability to other works and applicability in real time applications.

This presentation is part of Contributed Presentation “CP7 - Contributed session 7”

Authors:

Sameh Eisa (University of California, Irvine)

Keywords:

image compression, numerical linear algebra