Learn how search engines treat images and key image SEO strategies to implement with this comprehensive review.
The effect of images
Images are important because they communicate a lot of information quickly, just at first glance. An example is logos. An effective logo conveys a unique identity that is specific to a business. They can include a combination of words, letters, and graphic symbols that refer to products, services, industry sectors, or even company values.
Here are some examples of highly interactive images:
A standard spread is the center of a page of some websites. Images of products in electronic catalogs.
Avatars represent groups in social media. Graphical photo.
Plan. And ever fashionable infographics.
Image search has become one of the most common types of vertical search. Customers often search for images when they:
You are buying a product where aesthetics can be important.
You want to see what the article looks like. Let's consider the product.
Want to understand how to identify topics. Need to understand local or directory information.
Due to the high interest of consumers in images, search engines have been working for a long time with sophisticated methods to combine images with search terms. They weave images into keyword search results in what is often called "federated search" or "federated search." Google calls this "universal search."
Images in Search: How to Manage and Analyze Them
Images appear when you search for something in different SERP sections, such as:
Regular keyword search results, integrated image section (from global search).
Added thumbnails with rich snippets. Includes featured snippets.
Vertical image search results. News search results.
Video search results. It's instructive to know some of the ways Google and other search engines use images for user keyword searches.
Humans look at the pictures and quickly process them to understand the content based on years of experience of the natural world and biological image processing techniques. Computer systems, on the other hand, do not have these sophisticated image processing capabilities.
A strong basis of image search algorithm is the association of words with image content. This is accomplished by using various metadata associated with the image to create content.
Search engines use:
File name. An old story.
Text near images on web pages. Another text in the HTML of this image.
The link points to the image. And others.
But this kind of metadata is often scarce or unreliable. File name can be gobbledygook database ID.
Website users often leave subtitles and alt text.
Texts surrounding images and stories are a source of dirty information. For this reason, search engines work with algorithms to analyze images and images, comparing the content of images that are small in content metadata and images that have similar content and have rich text data.
Part of this algorithm analysis involves detecting text in images (ie symbols that appear in images) or text that artists have added to images. Optical character recognition (OCR) technology has been around for over a century, but it became more sophisticated when it was possible to convert scanned documents back to text documents starting in the 1970s and 1980s.
It is not known when OCR was first incorporated into the image scanning algorithms used by search engines, although it is likely that it was introduced around 2005. In 2005, Google helped make Tesseract an open source OCR service. They hired one of Tesseract's main developers, Ray Smith, who developed the technology at Hewlett-Packard. (Smith currently works in the research division of Alphabet's DeepMind, which is an AI thinking machine.)
In 2006, Google Labs launched the popular Labeler program, a crowdsourced way to identify images by asking people to submit keywords to describe the photos Google shows them.
Image Labeler supports the process by having two different users show the same image, and users will get points by matching the same word for the image. (Ten years later, Google brought this project back in a different form as part of its Crowdsource project.)
Image Labeler is believed to have generated a large amount of content data that has been aggregated to help train an AI-like algorithm to recognize features in images, improving Google image search.
Comments
Post a Comment