Video and Multimedia Technologies and Services Research

Video and multimedia content is now a part of one's daily life in many ways: television broadcasts, streamed online clips and movies, consumer videos of personal events, and even emails. As the amount of multimedia available continues to grow, the organization and discovery of this content increasingly needs automatic tools.

The Video and Multimedia Technologies and Services Research Department (part of the Networking & IP Services Lab within AT&T Labs Research) is continually investigating innovations that improve all interactions with video and multimedia, whether it's on a desktop PC, a mobile device, or an in-home television. To learn more about this work, click on a link to a project, either currently active and historical or one of the general areas below for more information.

Processing and Retrieval
logo_miracle_90_60 logo_imiracle_90_60 logo_vissem_90_60 logo_eclips_90_60
MIRACLE / CAE - Large-scale content acquisition and analysis engine framework.
iMIRACLE - Mobile retrieval and playback, models customized by metadata and content.
CAE Services - HTTP interfaces to processing, indexing, and retrieval.
Semantic Concepts - Machine learning and visual semantic classifiers for fast content filtering.
eClips - Personalized content alerts with relevance feedback for fast learning.
 
logo_ccd_90_60 logo_visbio_90_60
Content-Based Copy Detection - Discover content copies to enhance indexing and retrieval.
Visual Biometrics - Improve indexing with robust face- and lip-based recognition.
Real-time Multimedia Analysis - Enabling real-time audio-visual metadata streams in a small footprint.
Unsupervised Segmentation and Classification - Collate similar short-time event segments in unsupervised frameworks.
Standards - Incorporate innovations into new standards definitions.
Indexing and Representation
logo_consent_90_60 logo_cam_90_60 logo_socialtv_90_60
CONSENT - Identifying bitstream errors from delivery and decoding.
Social TV - Enhance your live TV experience with crowd intelligence.
CAM - Intelligent monitoring of multiple live metadata streams.
Summarization - Locate high-interest segments of video to identify unique content.
Delivery and Consumption