Super-resolution Capabilities of the Hough-Green Transform

Abstract

A novel approach to the Hough Transform (HT) calculation, based on tracing object contours on bitonal imagery, has recently been proposed. This approach, called the Hough-Green Transform (HGT), is typically much more computationally effective and accurate than the Standard HT (SHT). The additional potential of the HGT lies in its natural ability to compute the complete HT domain with sub-cell resolution. The paper focuses on achieving HT domain super-resolution, i.e., the derivation of a high-resolution HT domain from a downsampled spatial one, which enables significant acceleration of HGT evaluation. The related tradeoffs are also discussed.