Title :
A low-cost, real-time, hardware-based image demosaicking algorithm
Author :
Karloff, Anthony ; Muscedere, Roberto
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Windsor, Windsor, ON, Canada
Abstract :
This paper presents a real-time, highly cost-effective approach to image demosaicking targeted specifically for high-speed and compact hardware implementation. Many existing demosaicking methods require multiple iterations of processing and complex calculations which are not suitable for high throughput or compact applications. The proposed method focuses on simplifying the demosaicking process to reduce computational complexity, preserve final image quality and foremost reduce memory requirements by removing the need to buffer large areas of the image for processing. A hardware architecture for the proposed method is outlined and compared to other methods targeted for hardware realization with greatly reduced memory usage, improved image quality and speed.
Keywords :
CMOS image sensors; computational complexity; image segmentation; CMOS imaging sensors; computational complexity; hardware-based image demosaicking algorithm; image quality; memory requirements; multiple iterations; CMOS image sensors; Color; Computational complexity; Hardware; Image quality; Image reconstruction; Image sensors; Interpolation; Pixel; Sensor arrays;
Conference_Titel :
Electro/Information Technology, 2009. eit '09. IEEE International Conference on
Conference_Location :
Windsor, ON
Print_ISBN :
978-1-4244-3354-4
Electronic_ISBN :
978-1-4244-3355-1
DOI :
10.1109/EIT.2009.5189599