Presently, it is difficult to efficiently search for desired images on the Web. In order to solve this problem, Image Vortex is an image search engine implemented to solve this problem. Image Vortex is able to function in cases where the query image is unavailable, which is a problem for existing search engines. Since images are communicating with each other, images that are mutually similar get grouped together. The user is then able to see the entire image collection and locate the desired image more effectively.
ImageCruiser is an application built on top of technology developed by the Haseyama Laboratory. The application was developed by a group of Sapporo IT companies between 2007-2009 as part of the “Information Grand Voyage” project of the Ministry of Economy, Trade and Industry. Working on images without any tagged text information, ImageCruiser instantly extracts visual features, automatically clusters and lays out a large number of images. In this way, ImageCruiser performs grouping of similar images and videos without relying on textual data, and allows a user to intuitively find the desired image from a large image collection.
In order to find desired content on the Web, a user needs to provide keywords or some query media. EasyFinder estimates the user’s preferences from their past viewing and listening history, thus enabling the recommendation of new videos and music. Furthermore, EasyFinder implements searching for videos and music with a similar “feel”, as well as music with similar structure – tasks are difficult for conventional methods.
Video Vortex is a video search engine that works by iteratively selecting the “best” result to allow the user to incrementally narrow down the search results and find their desired video. Since, whenever we view a video, we consume various forms of information (visual, audio, semantic), Video Vortex implements a highly effective video search by utilizing the image, audio, music and text features.
Query is You (QIY)
QIY is a next-generation information retrieval system that recommends videos to its users. When you stand in front of a display, the system uses your facial expressions, movement and attire to estimate your interests and mood, and recommend matching videos. Imagine finding interesting TV programs or Internet videos just by sitting down in front of your TV or computer! We hope that the results of our research will make such things a reality.
Open Access Sharing Image System(OASIS)
OASIS a next-generation Social Network System (SNS) that allows people to communicate via images, using their smartphone or tablet PC. The system estimates people’s interests based on two forms of information: (1) preferred images that the user chose when shown a collection of images on their smartphone or tablet; and (2) the user’s physical activity obtained by the system while they were standing in front of the display. Next, when a user’s interests overlap with those of another user, the system can recommend a new image for the first user, sparking a new discovery.
On top of this, the Laboratory of Media Dynamics is also actively developing the the Tri-Media Vortex cross-media search engine and the Immersive Web video search engine.
Community-Oriented & Shared Multimedia recOmmendation System (COSMOS)
COSMOS is a next-generation of an information recommendation system that provides videos corresponding to users’ interest via communication between COSMOS and users. COSMOS estimates users’ interest from their behavior such as view and selection of information shown in the display; then, videos corresponding to users’ interest are recommended. Based on users’ repeated selection of information, COSMOS creates a new world of videos for each user.