Gray-Value-Based Encoding Representation Method for Images

In image template matching, although correlation algorithms based on pixel gray values are widely used, they suffer from high time complexity and sensitivity to brightness and size variations. To address these limitations, we propose a novel encoding representation method based on image gray values. This approach divides the image into fixed-size blocks (termed R-blocks), calculates the total gray value for each R-block, and encodes it based on its ordinal relationship with adjacent R-blocks. Template matching is achieved by comparing the encoded values of R-blocks. The computation of R-block encodings is simple, and matching only requires equality checks between codes, enabling fast comparison algorithms. The method demonstrates robustness to gray-level variations and noise.