The USC Andrew and Erna Viterbi School of Engineering USC Signal and Image Processing Institute USC Ming Hsieh Department of Electrical and Computer Engineering University of Southern California

Technical Report USC-SIPI-129

“On the Hardware Requirement for 2-D Image Convolution in Electro-Optical Systems”

by M. Eshaghian, D. K. Panda, and V. K. Prasanna Kumar

August 1988

In this paper, we study the hardware requirement for 2-D image convolution on an electro-optical model. We identify the basic components of an Optical model of parallel computation using optical interconnects and their relationship with well known VLSI grid model from a computational perspective. We show a lower bound of W(nw) on the volume requirement of any electro-optical chip for 2-D image convolution using a w x w window and a n x n image. This bound also applies to the area requirement of a VLSI chip. Designs available in the literature for computing the 2-D image convolution are compared with respect to their hardware requirement.

To download the report in PDF format click here: USC-SIPI-129.pdf (0.6Mb)