ClusterMedia Labs is a leading semantic audio-visual analysis
technology company based in Aveiro, Portugal (http://www.ua.pt/campus.asp - Building 1). The company
provides complete automatic metadata generation solutions for Media
Asset Management and Business Intelligence.
The research done is expressed in advanced algorithms for content-based
analysis. The innovative of technology allows: high accuracy rates
with low computational costs (real-time performance), even in
several and simultaneous semantic contexts (speakers, music background,
environment sounds); and a Metaphysical behavior (with more refined
semantic definitions of sound patterns).
ClusterMedia Labs has developed advanced algorithms to automatically
generate metadata from Live Broadcast 'Feeds' and audio-visual
archives, and publish online media content searchable in several
categories like speakers, TV shows, commercials, musical genres,
jingles and others. The company has launched in NAB 2008 (www.nabshow.com) and IBC
(www.ibc.org)
LiveMeans® Platform, an Automated Media Analysis for content-based and
semantic indexing, which enables real time notifications to be sent
regarding users personalized content on the mobile TV (avoiding users
from becoming frustrated with channel zapping and tailoring the
content, namely contextual advertising), IPTV and web TV services, as
well a novel and promising Content-Based Music Discovery &
Retrieval Engine.
ClusterMedia Labs’s intellectual property (IP) and products portfolio includes:
• LiveMeans Engine® -
adaptable platform that may be fitted to the desired application in the
following segments:
- Media Asset Management / Broadcast
- Media Monitoring/Clipping
- Parliaments
- Real-Time Content Notifications (e.g SMS Alerts)
- Automatic Start Recording for Live Mobile TV
- “SmarterYouTube” (Search Engine to improve youtube and similar
searches)
- Music On Radio
- Music On TV
• SoundsLike.Me is a
Content-Based Music Discovery & Retrieval Engine
The incorporation of this music acoustic similarity technology in
hybrid music recommendation systems in mobile music services will
introduce users to new exciting usages like to find songs and/or
artists (drag & search) easily by way of acoustic similarity with
recent material releases (novelty & relevance, improves current
discovery & search tools in music recommendation services and
mitigates obstacles related to scalability, "cold-start" and one-sized
selections and access to the long tail of unclassified music
collections like MySpace.
• TV & Radio Commercials Automatic
Detection