|
|
Robotics Institute
Seminar, November 4
Time
and Place | Seminar Abstract | Speaker
Biography | Speaker Appointments
Engineering Methodologies for Real-time Vision Systems
Visvanathan Ramesh
Siemens
|
Webcast |
|
Time and Place |
Special Location: NSH 3305
Refreshments
Talk
The proliferation of cheap sensors, faster computers and integrated networking has made a significant impact on the development of Computer Vision technologies of applicability to a variety of industrial sectors including: safety, security, automotive and industrial automation. The computer vision field has matured to a certain extent in its development of algorithms for sub-tasks such as: object detection, tracking, motion analysis, recognition, 3D from stereo, Structure from Motion, multi-view geometric reasoning, event analysis from video, etc. Trends in machine learning and pattern analysis techniques are also leading to better and robust classification methods, systems with adaptation and some degree of self-tuning. In addition, vision researchers have focused more on performance evaluation of total systems and in validation of algorithms on specific applications.
This talk will give an overview of "Computer Vision Systems Research" performed at Siemens Corporate Research with specific examples in safety, security and automation. We will, in particular, highlight our systems engineering methodology required to build robust, scalable and reusable systems and modules. We conclude with our view of long-term research needs in the field.
|
Speaker Biography |
For appointments, please contact Janice Brochetti
The Robotics Institute is part of the School of Computer Science, Carnegie Mellon University.