Visually appealing Image Recommendation based on Article Content using Artificial Intelligence

The media landscape has undergone a significant transformation in recent years due to technological advancements and the growing influence of social media. As a result, professionals in the industry, including journalists, content creators, authors, and news professionals, face constant challenges in this rapidly evolving environment. One of the most critical demands in the media industry is the need to publish content quickly on multiple platforms, incorporating both text and visuals, to effectively target audiences. Adding relevant images that capture the essence of the story and highlight the main keywords can significantly enhance readers’ interest and improve engagement. Thus, media professionals could adapt their publishing strategies to keep pace with changing trends and meet the evolving expectations of media consumers.

Given that there are several tools available in the market that support cross-posting, such as Planable, eClincher, Buffer, and Hootsuite, which can save time, and that professionals in the news and media industry have quick access to new information to fact-check against fake news, hoaxes, and scams (e.g., eufactcheck.eu, IFCN, Google fact check tools, FactCheck.org), the emphasis now falls on quickly pairing the article’s text with relevant image(s). This challenge requires developing image recommendations based on the text to create visually appealing stories, thereby saving time for news and media professionals, such as journalists. The AI4Media project has identified this challenge, which is aligned with the “C4-A: AI for suggesting visually appealing images based on text” challenge for AI use cases in news, reorganization, and content moderation.

The overall objective of VIREO is to develop an integrated digital solution that uses AI techniques to analyze the content (text) of an article and recommend a collection of images that would best accompany it in close-to-real-time. The purpose of this project is to benefit both authors (journalists) by allowing them to quickly select visually appealing images to create engaging and captivating stories, and consumers (readers) by providing an enhanced reading experience, increased engagement, and better recall.

The VIREO project ran for nine months (March 2023 - November 2023). For more information about the VIREO project, please contact us through the channels described in the official website of Human Opsis: humanopsis.com