Media Information Laboratory of NTT CS研 公式ホームページ

title_media_e.gif

icon_media.gifWe aim to implement functions that can identify and find people and objects by picking out sounds and voices from the vast amount of audio and video information that is obtained via microphones and cameras and made available by ubiquitous networks.

Robust media search

  • 04-2e.jpgFor media information such as video, audio and images, we are developing a technique for searching and identifying media content at high speed that is unaffected by signal distortion or noise. With this robust media search (RMS) technique, we can support the storage and distribution of information and so well handle the media information explosion.

Media information extraction

  • 05-1e.jpgWe are developing a technique for automatically extracting information related to objects and events included in media information such as video, audio and images. This technique connects media information with symbolic information such as text and attributes, thereby bringing new value to vast quantities of media information.

Audio/speech processing

  • 08-5e.jpgWith the world’s most advanced speech enhancement and speech recognition technology at its core, we are implementing technology that uses audio information to help us understand the surrounding environment as a communication scene. With this technology, we will support new services that help people to communicate.

Communication environment

  • TROOM2_S.jpgWe are researching technologies including conversation scene analysis using non-verbal information such as expressions and actions, conversation control aimed at using computer assistance to stimulate conversations, and a communication environment that supports collaborative work between multiple locations while giving users the same effectiveness as sharing the same room. We aim to create an interactive communication environment that can support the communication business of the near future.

Dynamical information processing

  • 10-1e.jpgBy making active use of disordered phenomena such as laser chaos, we are implementing a high-speed random number generator and an information-theoretic secure encryption system. With these technologies, we will contribute to the implementation of safe and secure communications.