Is SEO a content-based technique

Visual search

Author: Florian Kolbe

The development of the classic Google image search

The development of the classic image search by Google was initiated in 2001 by a photo of Jennifer Lopez in a green chiffon dress at the Grammy Awards. Shortly after this photo was published, the dress was searched for millions of times on Google. The search still counts today among the most popular searches of all time and was the starting signal for the development of Google's image search.

The Google image search is a context-based image search. This means that the Google bots analyze the context and surroundings of an image in order to determine its relevance for a search query. The graphic content of the image itself is not taken into account. Instead, the text of the web page that surrounds the image, the content of the Alt attribute that is linked to the image, the caption and also the file name provide the essential information that Google to categorize and sort images used. Providing this information is therefore an important part of image SEO. The information is indispensable for Google and other search engines in order to be able to understand with a high degree of accuracy what an image represents.

What is visual search and how does it differ from classic image search?

Illustration: Visual Search - Author: Seobility - License: CC BY-SA 4.0

Image search, in which a search engine performs a text-based query and tries to find the best visual match, is very different from visual search. Visual Search means Search engine users can search with images, not just after them. So in the visual search, an image can be used as a query instead of text. The term visual search therefore 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

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

  • Reverse Image Search
  • Related Search
  • Filtered Search & Deep Image Search
  • Augmented Reality Search

Reverse Image Search

With the reverse image search developed by Google it is possible to search the internet for a specific image. The image is used as a search query. Reverse image search was first used by companies trying to identify unauthorized uses of their product images. Reverse image search now also plays an important role in marketing. For example, often used stock photos can be found using the reverse image search. The list of results shows the websites that also use this image.

Related Search

Related Search was introduced by Pinterest. The technique, also known as content-based image search, is based on the Comparison between images or image content using visual computer techniques. These techniques make it possible to search for similar images or individual items in images. Related Search uses a logic similar to that of the Google Suggest function and generally provides results as an answer to a search query which, from the point of view of search engines, are similar to what is being searched for. The stock photo portal Shutterstock also uses a similar technology. For example, if you drag a photo of a group of young people or a meeting into the search mask, the search returns images on which these elements can also be seen.

Screenshot with example for Related Search from shutterstock.com

Filtered Search & Deep Image Search

The filtered image search also has its origin on Pinterest. This feature builds on the related search and suggests filters - like color and size - so that users can focus their search. For example, a lamp or a table can be marked on the picture of a stylish room. The search engine then delivers the appropriate products. The success of this visual search has led most search engines to adopt this search method.

The deep image search developed by Bing follows a similar approach. It enables the selection of objects in an image using a cropping tool and searching for related images and other information.

Augmented Reality Search

As part of the Augmented Reality Search, users can use the camera of their smartphone to use visual inputs for search queries. With the Lens apps from Google and Pinterest, users can photograph objects in reality and Call up related images and further information. For example, it is possible to take photos of sights and obtain background information about them. In everyday life, for example, a restaurant can also be photographed and Google provides information such as opening times and menu.

Despite these new possibilities, meta-search is still important in visual search. By combining context-based and visual search methods, the search results in the image search can be optimized.

How does visual search work?

The unconscious decisions that allow us to understand patterns in images have so far only been incompletely integrated into machines. However, in order to conduct an accurate visual search, search engines need a lot more sophisticated processes than traditional image searches. For this reason, artificial intelligence and neural networks are used in visual search to mimic how the human brain works when identifying image content.

The typical human process in which we decode the components of an image, filter out relevant information and discard unimportant information, the so-called distractors, is carried out by complex image recognition algorithms in visual search. Using these algorithms will be Color and brightness values ​​of the pixels, shapes and textures as well as other visual information extracted from an image, evaluated and compared with other image content. Based on this information, the search engines decide what an image is about and then conceptualize and categorize related elements.

This task is carried out in the visual search of neural networks. These networks operate without human intervention and change the way they operate based on feedback signals such as user behavior to deliver the desired output. The visual image search, like the voice search, would not be possible without artificial intelligence and machine learning algorithms.

Importance for marketing and SEO

Visual search offers many new opportunities for marketing. 93% of consumers say that they primarily consider visual factors when purchasing a product. With the help of image-based search queries, you can find exactly what you are looking for faster, thanks to visual search functions searching and shopping from a single image. As a result, shoppers who search with images will find a product they want faster and are less likely not to complete the purchase.

The visual search also opens up new opportunities for cross-selling. When users search for a specific product, there is a good chance that they will buy other products that they see in the same picture. For example, complete outfits can be presented in the area of ​​fashion and fashion, whereby the individual products can be searched using the deep or related search. IKEA has been doing this very successfully offline for years with the carefully coordinated living environments in its furniture stores.

Now and in the near future, marketers face little competition in visual search. In relative terms, visual search is still in its infancy and so far has rarely been part of marketing strategies. This presents a great opportunity for companies willing to invest in visual search results (VSERPs).

Visual search also brings about changes for search engine optimization. While classic image SEO is and will remain important, the product images that have become popular due to Amazon will have more and more obsolete against a white background in the future. For Visual search optimized images are high resolution and show the product in a natural setting. The user and potential customer thus get an impression of the benefits of the product in everyday life. The image optimization for the new image search therefore begins as soon as the images are taken.

Google, Bing and SEOs assume that Smartphones and other mobile devices will be used more and more as visual search engines in the future. For example, users take photos of a restaurant from the outside and, as a search result, receive detailed information about its offer and find out which friends have already visited the restaurant. It seems certain that visual search, alongside voice search, will make up by far the largest part of search queries in the near future. However, it has not yet been decided which of the search engines will win the race for the best visual search.

About the author
Florian Kolbe works for Online Solutions Group and works there in the area of ​​online marketing. During his work as a trainee, he deals extensively with the topics of SEO and SEA and always comes across new challenges and exciting topics. In this guest article he shares his knowledge about the new possibilities of visual search and their importance for marketing and search engine optimization.

Similar articles

To quote the article, just copy this link:
https://www.seobility.net/de/wiki/Visual_Search