Detecting the Most Unusual Part of a Digital Image

TL;DR

This paper presents an algorithm to identify the most unusual region in a digital image by measuring maximal distance to similar non-intersecting shapes, useful for analyzing large image datasets without predefined models.

Contribution

The paper introduces a novel algorithm for detecting the most unusual image part based on shape distance, applicable to large and unstructured image databases.

Findings

Effective detection of unusual image regions

Applicable to medical and large image databases

No prior model of interesting parts needed

Abstract

The purpose of this paper is to introduce an algorithm that can detect the most unusual part of a digital image. The most unusual part of a given shape is defined as a part of the image that has the maximal distance to all non intersecting shapes with the same form. The method can be used to scan image databases with no clear model of the interesting part or large image databases, as for example medical databases.