NewsInstaMiner: Enriching News Article Using Instagram

摘要

Instagram is one of the largest and most popular image sharing website. Millions of users share their real life events with the outside world by uploading images on Instagram. Most of these images are annotated with tags and captions which are high level descriptions of these images. So if a news event occurs, users might have captured the relevant images and share them on Instagram. We have observed that many images on Instagram are relevant to the popular News events. So we can recommend such images for the news articles related to the event. To the best of our knowledge, no existing work has been done yet to recommend images particularly for news articles. In this paper, we have developed an image recommendation technique namely NewsInstaMiner. This technique analyzes news articles to extract the high level concepts and based on these concepts it recommends images from Instagram. We have implemented NewsInstaMiner and used it to demonstrate the effectiveness of our proposed technique.