Querying Video Content at a Conceptual Level using Multi-modal features

Uma Srinivasan
CSIRO-MIS, Digital Media Information Systems

Tuesday 4 May at 11am

Abstract

A typical video sequence consists of both visual and auditory content and humans often rely on both modes of communication to recall a particular video sequence. While dealing with videos, the query and retrieval problem has been addressed in different ways by the database and image analysis groups, often using only one mode at a time. In this talk I will discuss our approach to querying video content at a conceptual level using both visual and audio features. I will also describe and demonstrate a prototype system that we have developed to query video content in the sports domain.

Short resume

Uma Srinivasan is a research scientist with CSIRO Mathematical and Information Sciences (CMIS), working in the area of Digital Media Information Systems where she has been focussing on multimedia data modelling and video information delivery. Her research activities have included heterogeneous datbase integration, knowledge discovery from databases, and complex data modelling. In her past life, Uma worked as a software consultant and Project director in the industry.

 

Back to HAIL Home Page