Abstract Discrete orthogonal polynomials are useful tools in digital image processing to extract visual object contours in different application contexts. This paper proposes an alternative method that extends beyond classic first-order differential operators, by using the properties of Krawtchouk orthogonal polynomials to achieve a first order differential operator. Therefore, smoothing of the image with a 2-D Gaussian filter is not necessary to regularize the ill-posed nature of differentiation. Experimentally, we provide simulation results which show that the proposed method achieves good performance in comparison with commonly used algorithms.
Rivero-Castillo, D., Pijeira, H., & Assunçao, P. (2015). Edge detection based on Krawtchouk polynomials. Journal of Computational and Applied Mathematics, 284, 244–250. https://doi.org/10.1016/j.cam.2014.11.017