Title :
Visualizing flows of images in social media
Author :
Itoh, M. ; Toyoda, Masashi ; Kamijo, T. ; Kitsuregawa, Masaru
Abstract :
Mass and social media provide flows of images for real world events. It is sometimes difficult to represent realities and impressions of events using only text. However, even a single photo might remind us complex events. Along with events in the real world, there are representative images, such as design of products and commercial pictures. We can therefore recognize changes in trends of people´s ideas, experiences, and interests through observing the flows of such representative images. This paper presents a novel 3D visualization system to explore temporal changes in trends using images associating with different topics, called Image Bricks. We show case studies using images extracted from our six-year blog archive. We first extract clusters of images as topics related to given keywords. We then visualize them on multiple timelines in a 3D space. Users can visually read stories of topics through exploring visualized images.
Keywords :
data visualisation; image retrieval; information retrieval systems; social networking (online); 3D visualization system; blog archive; change recognition; commercial picture design; image bricks; image cluster extraction; image flow visualization; mass media; product design; social media; Blogs; Feature extraction; Histograms; Market research; Media; Timing; Visualization;
Conference_Titel :
Visual Analytics Science and Technology (VAST), 2012 IEEE Conference on
Conference_Location :
Seattle, WA
Print_ISBN :
978-1-4673-4752-5
DOI :
10.1109/VAST.2012.6400539