|
|
|
Research Theme: Sensors & Mobility Until recently, computers and therefore our information systems were mostly deaf and blind. All information about the real physical or social world had to be manually entered into these systems. In the last several years, however, this situation began to change. First, enterprise systems are routinely capturing data reflecting many of the central processes of the business. Second, many kinds of physical sensors have become available and are being integrated with business applications. Third, the Web has become a very rich source of information for businesses. Each source requires different types of “sensors” – data analysis for business process data, physical sensors to report on physical world, and text processing for unstructured documents. In reality, the sensors of all three kinds provide only partial coverage of the areas of business interest and in many cases the collected information is probabilistic and noisy. The focus of our research in this area is the collection, interpretation and integration of heterogeneous and noisy information from multiple sources. Projects Some of our recent projects include:
People
A Bayesian Framework for
Robust Reasoning from Sensor Networks Multiple-Camera People
Localization in a Cluttered Environment Adaptive Algorithms for
Pitch-synchronous Speech Signal Segmentation Developing Audio Processing Agents
for Multi-Agent Mpeg-7 Enabled Environment The Community of Multimedia
Agents Pitch-synchronous Speech
Signal Segmentation and its Applications Indexing and Retrieval of Music via
Gaussian Mixture Models
|
|
|
