Introduction

An image hash can be considered as a cursory summary of an image. You can use image hash analysis to identify exact duplicates in a set of images, because identical images produce identical hashes. Image hash analysis can tolerate small distortions, but an image that is subject to significant changes results in a different hash.

Image hash analysis is much faster than object recognition or image comparison because it considers each image as a whole and does not analyze the detail within an image. To identify which parts of an image have changed, use image comparison instead. If you need a solution that can identify matching images while tolerating distortion (including scaling, rotation, and perspective changes or skew), consider using object recognition.