Finding copies of an image: a comparison of reverse image search systems on the WWW

Objective - The reported investigation aimed to assess the performance of services that offer search by image on the Internet / WWW to find copies of a known image. The motivation is that finding copies of an image can be useful for several purposes, while search by image is a relatively new method that allows finding such images and that is offered free of charge by a few Internet search services. Furthermore, the context of each revealed image copy may yield relevant information. Methods - Various images have been used in empirical case studies. Results - Finding copies is positioned in a structured view of applications of reverse image searching. The concept 'copy' is clarified and sharpened. Data have been collected regarding the performance of several contemporary services to find copies of an image, namely TinEye, Google and Yandex. Conclusions - The reported findings have revealed significant differences in the number of copies discovered and in the precision of the search results. Therefore, all this may be helpful for users / practitioners to select andn appropriate image search service to reveal copies of an image and even information that is related to the image.