Reverse Image Search

Definition

Reverse Image Search (also called “reverse image lookup”) are tools that query and identify instances of the exact image or similar images found on publicly indexed websites based on the original image’s attributes and underlying metadata.

Because these tools report on instances of use/reuse, they are frequently used by copyright holders (including artists, photographers, and corporations) to investigate unauthorized uses of images over the web.

Practitioners can use reverse image search to locate digital objects from their collections’ external websites, including some social media outlets and blogs.

Applications for assessing digital content use/reuse

Reverse image search follows a process similar to a typical text search engine. It requires the user to select an image (from a source file, user sketched digital image or URL) and submit it to the reverse image search engine. The engine then scans publicly indexed websites and returns a list of websites that contain an exact image or adjusted image (including cropped, resized, rotated, or color adjusted versions of the queried image) matches. Reverse image search results are displayed very similarly to other types of search results. They typically include the title of the webpage, URL, image thumbnail, image resolution size, and a brief summary of the website.

Reverse image search offers practitioners the ability to locate instances of images from their collections being used/reused across the web. This is particularly helpful because these search engines can query thousands of websites, including those that a practitioner might overlook.

Practitioners can use these results as data for assessing use/reuse. Several case studies (see resources below) provide examples of practitioners querying a reverse image search engine with images from their collections, compiling results, developing an analytic framework to evaluate the websites, and reviewing URLs that contain the image queried to understand:

Where users are posting images
The types of users who are posting images
The purpose(s) and/or motivation(s) for posting images
Any information or comments that the posted image might have generated
The overall frequency of use/reuse instances

Tools

Evolution of the tools

Reverse image search tools have developed over time. Conceptualized in 1992, reverse image search was intended to search discrete databases for images using the attributes of the image such as pixels, color, shape, and texture and retrieve the exact or similar image. As computer, imaging, and internet technology advanced, the underlying search algorithms changed. Today’s reverse image search doesn’t look or act the same as it did even 5 years ago. Machine learning, massive search results, and huge image databases have changed the way these tools search and what results are returned.

In earlier versions of the tool, reverse image search only used the images attributes as the search query. Today most image search tools use not just the image but the underlying metadata, subject type headings, and machine learned past searches for the same image as the search query. How the tool searches depends on the tool algorithm. All search engines will augment the search using a text string.

Commonly used tools include:

Ethical guidelines

Practitioners should follow the practices laid out in the “Ethical considerations and guidelines for the assessment of use and reuse of digital content.” The Guidelines are meant both to inform practitioners in their decision-making, and to model for users what they can expect from those who steward digital collections.

Strengths

Reverse image search provides practitioners with a way to locate instances of reuse/sharing of digital objects within indexed web spaces.
Reverse image search platforms have lower barriers of use than using search terms to discover reuse/sharing.
The search results from the largest of the reverse image search platforms (Google Images and TinEye) will return more accurate results due to their massive image databases.

Weaknesses

Reverse image search is not always 100% accurate. This process may return false positive results.
Reverse image search results are based on the number of times an image has been searched and returned. Unpopular images will either not be found or be relegated to lower search result pages.
Reverse image search results are dynamic. As databases grow and more searches are performed, these websites will display different search results. Clearing a browser’s history will also change search results.
Reverse image search cannot identify the reuse of images on websites that are not publicly indexed or unavailable to search engines crawlers such as Google or Bing.
Reverse image search cannot access content outside of HTML files. PDF documents, audio, and video files placed on websites are not queried by reverse image search.
Reverse image search can only query static images on the web, although multiple GLAMR case studies have discussed how these engines can match images that have been integrated into videos.
Poor image quality may prevent some results from being identified by search crawlers.
Many popular reverse image search engines do not provide documentation on how their search algorithms operate. They also do not articulate what information they use or collect in order to run the searches.
Many reverse image search engines do not allow for batch searching. Practitioners must complete a manual search, one-at-a-time. This can be time consuming and suggests that this approach would not scale to assess a practitioner’s entire digital collections. One alternative to this is TinEye MatchEngine, which is a fee-based service that offers batch querying.
Practitioners should note that reverse image search alone will NOT provide the analysis needed to perform a use/reuse assessment. Instead, reverse image search can generate a list of data points, which then the practitioner will need to explore further, using a variety of possible methods, to assess the presence and extent of use/reuse.
Practitioners can only evaluate use/reuse with the information they can obtain from reviewing websites. Frequently, this approach will not provide practitioners with all of the data they may need to appropriately understand users or their motivations.

Learn how practitioners have used this method

Case Study: Assessing the reuse of an academic library digital collection using reverse image search
Kelly investigated the effectiveness of two reverse image search engines, Google Image Search and TinEye, by querying digital objects found in a local digital library collection. The author had only limited success locating instances of digital objects posted across the web. As a result, Kelly concluded that practitioners should be aware of the resources needed and limitations of the search engines when making a determination to use reverse image search in assessing digital objects.

Kelly, E. J. (2015). Reverse image lookup of a small academic library digital collection. Codex: the Journal of the Louisiana Chapter of the ACRL, 3(2), 80-92.
Case Study: Assessing the reuse of a museum digital collection using reverse image search
Kirton and Terras leveraged reverse image search as one part of their study to understand how images from the National Gallery in London, UK were being shared across the web. Utilizing Google Image Search and TinEye search engines, Kirton and Terras generated a data set of over 3,000 instances of digital objects being posted on the web. They conducted qualitative analysis on the data and found that image use/reuse was driven by its subject matter. They also “triangulated” reverse image search findings with web analytics data to develop a more nuanced understanding of sharing patterns, noting that more frequently visited pages often had more frequently shared images across the web.

Kirton, I., & Terras, M. (2013). Where Do Images of Art Go Once They Go Online? A Reverse Image Lookup Study to Assess the Dissemination of Digitized Cultural Heritage. In Museums and the Web 2013, N. Proctor & R. Cherry (eds). Silver Spring, MD: Museums and the Web.

Additional resources

Beaudoin, J. E. (2016). Content-based image retrieval methods and professional image users. Journal of the Association for information science and technology, 67(2), 350-365.

Kelly, E. J. (2015). Reverse image lookup of a small academic library digital collection. Codex: the Journal of the Louisiana Chapter of the ACRL, 3(2), 80-92.

Kirton, I., & Terras, M. (2013). Where Do Images of Art Go Once They Go Online? A Reverse Image Lookup Study to Assess the Dissemination of Digitized Cultural Heritage. In Museums and the Web 2013, N. Proctor & R. Cherry (eds). Silver Spring, MD: Museums and the Web.

Kousha, K., Thelwall, M., & Rezaie, S. (2010). Can the Impact of Scholarly Images Be Assessed Online? An Exploratory Study Using Image Identification Technology. Journal of the American Society for Information Science and Technology, 61(9), 1734–1744.

Nieuwenhuysen, P. (2019). Finding copies of an image: a comparison of reverse image search systems on the WWW. In Collaboration–Impact on Productivity and Innovation: Proceedings of 14th International Conference on Webometrics, Informetrics and Scientometrics & 19th COLLNET Meeting 2018, 97-106.

Nieuwenhuysen, P. (2018). Image search process in the Web using image copy. Journal of Multimedia Processing and Technologies, 9(4), 124-133.

Reilly, M. (2021). Digital Image Users and Reuse: Enhancing practitioner discoverability of digital library reuse based on user file naming behavior (Doctoral dissertation).

Reilly, M., & Thompson, S. (2017). Reverse Image Lookup: Assessing Digital Library Users and Reuses. Journal of Web Librarianship, 11(1), 56–68.

Reilly, M. (2021). Digital Image Users and Reuse: Enhancing practitioner discoverability of digital library reuse based on user file naming behavior (Doctoral dissertation).

Thompson, S., & Reilly, M. (2017). A Picture Is Worth a Thousand Words’: Reverse Image Lookup and Digital Library Assessment. Journal of the Association for Information Science and Technology, 68(9), 2264–2266.

Contributors

Contributors to this page include Santi Thompson and Michele Reilly.

Cite this page

Reilly, M., Thompson, S. (2023). Reverse Image Search. Digital Content Reuse Assessment Framework Toolkit (D-CRAFT); Council on Library & Information Resources. https://reuse.diglib.org/toolkit/reverse-image-search/

Reverse image search tutorial