Reflection Invariance: an important consideration of image orientation

TL;DR

This paper emphasizes the importance of reflection invariance in computer vision, highlighting its absence causes inconsistencies in current systems and advocating for its consideration as a quality measure.

Contribution

It introduces reflection invariance as a crucial property, demonstrates its impact on object detection and scene classification, and calls for its integration in future algorithm assessments.

Findings

Inconsistencies in object detection with mirrored images

Scene classification errors due to lack of reflection invariance

Some feature detection methods exhibit partial invariance

Abstract

In this position paper, we consider the state of computer vision research with respect to invariance to the horizontal orientation of an image -- what we term reflection invariance. We describe why we consider reflection invariance to be an important property and provide evidence where the absence of this invariance produces surprising inconsistencies in state-of-the-art systems. We demonstrate inconsistencies in methods of object detection and scene classification when they are presented with images and the horizontal mirror of those images. Finally, we examine where some of the invariance is exhibited in feature detection and descriptors, and make a case for future consideration of reflection invariance as a measure of quality in computer vision algorithms.