Combining Signature Method and Linear Filtering for Convex Polygon Detection in Slanted Images

Abstract

This paper presents a novel algorithm that incorporates signature method and linear filtering techniques to achieve convex polygon detection in slanted or deformed input image. The algorithm is shown applicable to effectively locating finder and alignment patterns essential for correctly carrying out the subsequent QR-code decoding process. Because only a single scanning is needed, the algorithm is computationally efficient and suitable for real-time applications.