Histogram computation using reconfigurable memory

Abstract

When it comes to image processing systems histogram  computation  plays  a   very   Vital   role.   Some   of  its applications are Image-Registration(IR), Show-through- correction method in an image which is a scanned printed document in duplex-mode, etc. When it comes to intensity based multimodality image-registration mutual-information(MI) has turned out to be one of the best metric. We need to obtain both individual histogram as well as joint-histogram for computation of mutual-information. With increase in histogram size the demand for Hardware resources for histogram computation increases drastically. This paper presents an energy efficient clock-gated memory-based parallel-algorithm for histogram computation. The proposed method manages to utilise less hardware,where the hardware usage is reduced by 27.03% compared to the other available literatures and as well as consume less power,where it consumes 4 mW power, ending up reducing the power by 50%.