View source for Visual Search

<seo title="Visual Search" metadescription="Learn the difference between classic image search and visual search as well as its functioning and importance for SEO."/>

== The development of Google’s classic image search ==

The development of Google's classic image search was initiated in 2001 by a photograph of Jennifer Lopez in a green chiffon dress at the Grammy Awards. Shortly after this photo was published, Google received millions of search queries for the dress. It is still one of the <strong>most popular search queries of all time</strong> and was the kickoff for the development of Google's image search.
Since its beginnings, Google’s image search has been a <strong>context-based search engine</strong>. This means that [[Googlebot]] analyzes the context and environment of a picture to determine its relevance for a particular search query. The graphic content of the image itself is not taken into account. Instead, the text surrounding it, the content of its [[ALT Attributes|alt attribute]], the caption, and also the filename provide the key clues Google uses to <strong>categorize and sort images</strong>. Providing this information is an important part of [[Image SEO|image SEO]] because it is essential for Google and other search engines to accurately understand what a picture shows.

== What is visual search and what is the difference to classical image search? ==

[[File:Visual-Search.png|thumb|450px|right|alt=Visual Search|'''Figure:''' Visual Search - Author: Seobility - License - [[Creative Commons License BY-SA 4.0|CC BY-SA 4.0]]|link=https://www.seobility.net/en/wiki/images/9/9d/Visual-Search.png]]

Google’s image search takes a text-based query and tries to find the best visual match. Visual Search, on the other hand, allows you <strong>to search with images</strong>, not just for them. This means that instead of text, you can use an image as a query.

Visual search refers to visual data entry and retrieval, including the new reverse image search technology and the traditional keyword-in/image-out model.	

== Types of visual search ==

Visual search is based on the recognition of objects and the comparison of visual information with known image content. There are a number of different types of visual search that use different techniques to identify and compare the content of images. The types of visual search are:

*Reverse image search
*Related search
*Filtered search & deep image search
*Augmented reality search

=== Reverse image search ===

Google’s reverse image search allows you to <strong>search the internet for a specific picture</strong> by using your image as a search query.  The first users of reverse image search were companies, attempting to identify unauthorized use of their product photos. Today, reverse image search is also important for marketing. Frequently used stock photos, as an example, can be found with the help of reverse image search. The list of results shows the websites that also use this image.

=== Related Search ===

Pinterest introduced related search, which is also known as content-based image search. It <strong>compares images or image content utilizing visual computer techniques</strong>. This method makes it possible to search for similar images or individual objects in images. Related search thus uses logic similar to Google's suggest feature and delivers results that are similar to what is being searched for. The stock photo provider Shutterstock uses similar technology. For example, if you drag a photo of a group of young people or a meeting into the search box, it provides images that also show these elements.

[[File:Shutterstock related search example.png|link|border=|700px|alt=Related Search as type of visual search|Screenshot showing related search]]

Example showing related search from [https://www.shutterstock.com/ shutterstock.com]

=== Filtered search & deep image search ===

The filtered image search also originates from Pinterest. This feature builds on related search and suggests <strong>filters - such as color and size - </strong>so you can narrow your search. For example, a lamp or table can be selected on the image of a stylish room. The search engine then delivers matching products. Most search engines adopted this method due to its success.

The deep image search developed by Bing follows a similar approach. It lets you <strong>select objects in an image using a clipping tool</strong> and finds related images and other information.

=== Augmented reality search ===

Augmented reality search allows you to use your smartphone camera to enter visual information as search queries. You can <strong>take pictures of objects and retrieve related images and information</strong> using Google's and Pinterest's lens apps. In everyday life, you could use this type of search with pictures of sights to get background information. You can also shoot a picture of a restaurant to retrieve its opening hours and menu. 

Despite these new possibilities, meta search is still important for visual search. The combination of context-based and visual search methods can optimize the search results of image search.

== How does visual search work? ==

The unconscious decisions that allow us to understand patterns in images can only be partially integrated into machines so far. However, in order to perform an accurate visual search, search engines need much more <strong>sophisticated processes than traditional image searches</strong>. For this reason, artificial intelligence and neural networks are used in visual search to mimic the way the human brain works in identifying image content.

In visual search, complex image recognition algorithms take over the typical human process in which we decipher the components of an image, filter out relevant information and discard irrelevant distractors. These <strong>algorithms extract</strong> color and brightness values of pixels, shapes, and textures as well as other visual <strong>information from an image</strong>, evaluate them and compare them with other image content. Based on this information, search engines decide what an image is about and then conceptualize and categorize related elements.

In visual search, this task is performed by neural networks. They operate without human intervention and change their functionality based on feedback [[User Signals|signals]] such as user behavior to deliver the desired output. Visual image search, as well as voice search, would not be possible without artificial intelligence and machine learning algorithms.

== Importance for marketing and SEO ==

For marketing, visual search offers many new opportunities. 93% of consumers state that they consider visual factors when buying a product. With image-based search, they can quickly find just what they are looking for because they can <strong>search and shop with a single image</strong>. Faster results also increase the likelihood of purchases being completed. 

In addition, visual search opens up <strong>new cross-selling opportunities</strong>. When users look for particular products, they are more likely to buy other goods they see in the same image. For example, complete outfits can be presented in the fashion sector, where individual items can be searched for using deep or related search. An offline example of this technique would be IKEA, with carefully coordinated living worlds in its furniture stores.

Currently and in the near future, marketers have little competition when it comes to visual search. Visual search is still in its infancy and has rarely been part of marketing strategies. This offers a great opportunity for companies willing to invest in visual search results (VSERPs).

Visual search also changes search engine optimization. While classic image SEO is still important and will remain so, the white background product images made popular through Amazon will become more and more obsolete in the future. <strong>Images optimized for visual search</strong> are high-resolution and show products in a natural environment. Users and potential customers thus get an impression of the product's usefulness in everyday life. Image optimization for visual search thus begins with the creation of suitable pictures.

Google, Bing and SEO experts believe that <strong>smartphones and other mobile devices</strong> will increasingly be used as visual search engines in the future. For example, users can photograph a restaurant from the outside and get detailed information about what it has to offer and which friends have already visited it. It seems certain that visual search will make up the largest proportion of search queries in the near future, alongside voice search. However, it has not yet been decided which of the search engines will win the race for the best visual search.

== Related links ==

*blog.adobe.com/en/2018/12/12/see-it-search-it-shop-it-how-ai-is-powering-visual-search.html
*https://www.socialmediatoday.com/news/why-visual-search-will-be-one-of-the-biggest-digital-marketing-trends-of-20/545999/

== Similar articles ==

* [[Image SEO]]
* [[ALT Attributes]]

[[Category:Search Engine Optimization]]

<html><script type="application/ld+json">
    {
      "@context": "https://schema.org/",
      "@type": "ImageObject",
      "contentUrl": "https://www.seobility.net/en/wiki/images/9/9d/Visual-Search.png",
      "license": "https://creativecommons.org/licenses/by-sa/4.0/",
      "acquireLicensePage": "https://www.seobility.net/en/wiki/Creative_Commons_License_BY-SA_4.0"
    }
    </script></html>