MENLO PARK, Calif. Feb. 9, 2011 Ajay Divakaran Andreas Stolcke
IEEE is the world’s largest technical professional association dedicated to advancing technology for the benefit of humanity in the areas of electrical and computer sciences, engineering and related disciplines. The Fellow distinction is IEEE’s highest membership grade and conferred by the Board of Directors to select members for their extraordinary record of accomplishments in any of the IEEE fields of interest.
Divakaran was honored for his contributions to multimedia content analysis. This work enables the automatic indexing of broadcast video allowing users to pull snippets of content of interest. He has wide ranging expertise in analysis of diverse audio-visual content for understanding and locating key events in video and audio.
Divakaran currently works on vision technologies, focusing on people, vehicle and vessel tracking and fingerprinting, as well as audio analysis. He led the development of SRI Sarnoff’s ACT-Vision product, which is the world’s first commercialized multi-camera hand-off and tracking system. He also led the development of SRI Sarnoff’s vehicle and vessel fingerprinting systems. He is currently working on technology to track people in dense crowds, multi-modal training systems, automatic food identification and volume assessment, classification of audio captured in varying environments, and audio source separation.
Stolcke was honored for his contributions in statistical language modeling, automatic speech recognition and understanding, and automatic speaker recognition. He has experience in detailed statistical models of speech and natural language, combining modeling at the level of words, prosody, pronunciation, and acoustics for more accurate recognition of speech, speaker identity, and speaker intent.
At SRI, Stolcke has been involved in key speech technology projects such as recognition and understanding of multi-party meetings for DARPA’s Personalized Assisted that Learns (PAL) program, and new feature extraction and modeling techniques for SRI’s speaker recognition system. He also designed SRI Language Modeling toolkit (SRILM), an open-source toolkit for statistical language modeling that is widely used in the research community.
About SRI International
SOURCE SRI International