A Method for Image Reduction Based on a Generalization of Ordered Weighted Averaging Functions

TL;DR

This paper introduces Dynamic Ordered Weighted Averaging Functions (DYOWAs), a new class of aggregation operators that generalize existing functions like min, max, and mean, for improved image reduction and noise reduction.

Contribution

The paper proposes DYOWAs, a novel generalization of OWA functions with input-dependent weights, applicable to image processing tasks such as reduction and noise filtering.

Findings

DYOWAs can effectively reduce images and noise.

Application of DYOWAs improves image quality.

Operators generalize common aggregation functions.

Abstract

In this paper we propose a special type of aggregation function which generalizes the notion of Ordered Weighted Averaging Function - OWA. The resulting functions are called Dynamic Ordered Weighted Averaging Functions --- DYOWAs. This generalization will be developed in such way that the weight vectors are variables depending on the input vector. Particularly, this operators generalize the aggregation functions: Minimum, Maximum, Arithmetic Mean, Median, etc, which are extensively used in image processing. In this field of research two problems are considered: The determination of methods to reduce images and the construction of techniques which provide noise reduction. The operators described here are able to be used in both cases. In terms of image reduction we apply the methodology provided by Patermain et al. We use the noise reduction operators obtained here to treat the images obtained in the first part of the paper, thus obtaining images with better quality.