Swift template matching based on equivalent histogram

Histogram-based template matching is an important method to search the globe optimization exhaustively. However, this method is commonly algorithmic complex. In this paper, we propose to replace the traditional histogram-based method with equivalent histogram-based method, which distinctly improves the matching efficiency. We first introduce the equivalent histogram on the basis of the relative centralization of the template's color information and prove the equivalence. Then, we discuss the application of equivalent histogram in the current algorithms and analyze the algorithmic complexity. The equivalent histogram calculates the histograms and their distances according to the relative centralization of color information, which decreases the memory and computation spending from the calculation of redundant information. The experimental results indicate that equivalent histogram-based method remarkably improves the matching efficiency with no degradation of matching effect.