Asynchronous Regional Computation Capabilities for Digital Retinas

Author

Gies, Valentin ; Bernard, Thierry M. ; Mérigot, Alain

Author_Institution

ENSTA, Paris

fYear

2006

Firstpage

7

Lastpage

11

Abstract

Efficient regional computation is a key for low power mid-level image processing. In this paper, we show how to fit digital retinas with such capabilities. We propose a low power and fast hardware software solution to perform regional integral operations based on a regional sum primitive. It takes advantage of a mixed synchronous-asynchronous approach and specific computation algorithms. To illustrate its behavior, a case study on over-segmentation reduction, using regional statistical methods, is presented.

Keywords

CMOS image sensors; computer vision; digital signal processing chips; image resolution; image segmentation; statistical analysis; asynchronous regional computation capabilities; digital retinas; hardware software solution; low-power mid-level image processing; over-segmentation reduction; pixellic processor; regional statistical methods; Adders; CMOS technology; Computer aided instruction; Computer architecture; Computer networks; Computer vision; Image analysis; Image processing; Machine vision; Retina;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer Architecture for Machine Perception and Sensing, 2006. CAMP 2006. International Workshop on

Conference_Location

Montreal, Que.

Print_ISBN

978-1-4244-0685-2

Electronic_ISBN

978-1-4244-0686-9

Type

conf

DOI

10.1109/CAMP.2007.4350341

Filename

4350341