A Novel Image Transformation For Solving Complex Image Mapping Problems

Abstract

The paper proposes a novel image transformation called Image Local Response (ILR) that can be used for solving complex image mapping problems. The proposed transformation brings together two approaches based on the pixel value distribution and image features. Image local response is defined as the average value of the difference between the transformed and the original copies of the same image whereby the transformation is small, i.e., the components of the corresponding parameter vector have sufficiently small unit values. The response has a few interesting properties useful in image mapping. The validity of the proposed image transformation is shown on sample complex image mapping problems formulated as the multi-objective piecewise imaging optimization problem.