Low-cost implementation of a super-resolution algorithm for real-time video applications

In this paper, a novel algorithm based on superresolution (SR) techniques for increasing the quality of video sequences, is presented together with its mapping into the Philips Research proprietary hardware-software platform. As result, a low-cost real-time implementation is obtained, suitable for personal multimedia applications. Low cost constraints are accomplished by re-using a video codec: and introducing some changes in order to avoid the use of specific SR hardware. As a consequence, a drastic reduction in the memory requirements is obtained at the expense of less than a 7% quality loss with respect to previous works.

[1]  Kiyoharu Aizawa,et al.  Very high resolution imaging scheme with multiple different-aperture cameras , 1993, Signal Process. Image Commun..

[2]  Cris L. Luengo Hendriks,et al.  Improving Spatial Resolution in Exchange of Temporal Resolution in Aliased Image Sequences , 1999 .

[3]  Gustavo Marrero Callicó,et al.  Low-cost and real-time super-resolution over a video encoder IP , 2003, Fourth International Symposium on Quality Electronic Design, 2003. Proceedings..

[4]  Jim Lin,et al.  HW-SW co-design and verification of a multi-standard video and image codec , 2001, Proceedings of the IEEE 2001. 2nd International Symposium on Quality Electronic Design.

[5]  Daniel Gross,et al.  Improved resolution from subpixel shifted pictures , 1992, CVGIP Graph. Model. Image Process..

[6]  R. Sethuraman,et al.  A low-cost implementation of super-resolution based on a video encoder , 2002, IEEE 2002 28th Annual Conference of the Industrial Electronics Society. IECON 02.

[7]  Gustavo M. Callico,et al.  Mapping of real-time and low-cost super-resolution algorithms onto a hybrid video encoder , 2003, SPIE Microtechnologies.