A new radiomics feature: image frequency analysis

TL;DR

This paper introduces a novel radiomics feature based on image frequency analysis that quantifies the relationship between lesions and surrounding tissue by analyzing value changes across image rows and columns.

Contribution

It proposes a new radiomics feature that captures lesion-surrounding tissue interactions through image frequency analysis, enhancing image characterization.

Findings

New feature effectively captures lesion-tissue interrelation

Improves lesion characterization accuracy

Potential for better diagnostic insights

Abstract

Radiomics is a promising technology that focuses on improvements of image analysis, using an automated high-throughput extraction of quantitative features. However, the character of lesion is affected by the surrounding tissue. A lesion on medical image should be characterized from the inter-relation between lesion and surrounding tissue as well as property of the lesion itself. The aim of this study is to introduce a new radiomics feature which quantitatively analyze the inter-relation between lesion and surrounding tissue focusing on the value change of rows and columns in a medical image.