Image Relevance on Websites and Readability

Abstract

AbstractThe World Wide Web has become a major source of information for acquiring knowledge and information. People hardly have time to sit down and read anything lengthy anymore so graphical contents are sometimes more effective for the readers than just the written text. However, irrelevant graphical contents on the web equally contribute to poor readability, distracting the reader from the focus of the reading. The main objective of this paper is to help web designers and developers to construct better web pages from a readability point of view. We propose a new methodology to measure the relevancy of text images on a webpage based on their similarity with the webpage text. The methodology combines different techniques to extract text from images and read text from web pages in order to find relevancy between them. This approach was used to analyze 50 different educational websites in Pakistan to automatically find the relevancy of their image. Our results indicate that the images which are irrelevant to the context of the page and poor-quality images cause lower relevancy scores. Thanks to this study, web designers can improve the readability of their web pages by modifying the graphical content according to the recommendations done.KeywordsReadabilityGraphical contentExtracted textRelevancyWeb page