BranislavKisacanin2011

Chair for Computer Aided Medical Procedures & Augmented Reality
Lehrstuhl für Informatikanwendungen in der Medizin & Augmented Reality

THIS WEBPAGE IS DEPRECATED - please visit our new website

Invited talk by Dr. Branislav Kisacanin

Branislav.jpg

Dr. Branislav Kisacanin
Vision, Video, and Imaging Lab Systems and Applications R&D Center Texas Instruments, Inc.

  • Date: Friday, 23 September 2011
  • Time: 3pm
  • Location: MI 03.13.010

The Embedded Vision Approach to the Development of Backover Prevention Prototype

Abstract:

In this talk we describe the development of a prototype for Backover Prevention on an embedded processor. The approach is based on developing (1) memory-agnostic compute functions and (2) optimal data transfers, independent of each other. Thus the data transfer design is orthogonal to the design of vision functions. This feature of our approach to embedded vision significantly simplifies the often very hard process of implementing complex vision algorithms on embedded processors.

Short Bio:

Dr. Branislav Kisacanin is a computer scientist with Texas Instruments Vision R&D group. He is primarily involved in vision algorithm and architecture design. Dr.Kisacanin has co-authored and co-edited several books, most recently "Embedded Computer Vision" (Springer 2008), and has co-organized a number of workshops on embedded vision and human-computer interaction at major vision conferences, currently organizing an HCI workshop at 2011 IEEE ICCV in Barcelona. In his spare time Dr.Kisacanin likes to show his kids the fun side of science and technology. Most recently they made a Tetris-playing NXT robot: http://www.youtube.com/watch?v=wY83EaE7svA (or search for "Tetris Bot" on YouTube?).

WebEventForm
Title: The Embedded Vision Approach to the Development of Backover Prevention Prototype
Date: 23 September 2011
Location: MI 03.13.010
Abstract: In this talk we describe the development of a prototype for Backover Prevention on an embedded processor. The approach is based on developing (1) memory-agnostic compute functions and (2) optimal data transfers, independent of each other. Thus the data transfer design is orthogonal to the design of vision functions. This feature of our approach to embedded vision significantly simplifies the often very hard process of implementing complex vision algorithms on embedded processors.
Imageurl: Branislav.jpg
Type: News
Videourl:  
Conferencelink:  


Edit | Attach | Refresh | Diffs | More | Revision r1.2 - 21 Sep 2011 - 14:59 - SlobodanIlic

Lehrstuhl für Computer Aided Medical Procedures & Augmented Reality    rss.gif