The World Wide Web provides an increasingly powerful and popular publication mechanism. Web documents often contain a large number of images serving various different purposes. This paper focuses on images that are associated with a story or preview to a story. Such images often accompany the key content on a web page, thus their identification is important for applications such as web page summarization and mobile access. We present a novel algorithm for automatic identification of story/preview images which combines features extracted from both the image itself and the surrounding text. The effectiveness of this algorithm is demonstrated by experimental results on over 1500 images collected from 25 news web sites.