SHUVRA S. BHATTACHARYYA
is a Professor in the Department of Electrical and Computer Engineering, University of Maryland at College Park. He holds a joint appointment in the University of Maryland Institute for Advanced Computer Studies (UMIACS). He is coauthor or coeditor of six books and the author or coauthor of more than 150 refereed technical articles. His research interests center around architectures, methodologies, software techniques, and tools for design of signal processing systems. He received the B.S. degree from the University of Wisconsin at Madison, and the M.S. and Ph.D. degrees from the University of California at Berkeley. He has held industrial positions as a Researcher at the Hitachi America Semiconductor Research Laboratory (San Jose, California), and Compiler Developer at Kuck & Associates (Champaign, Illinois). He has held a visiting research position at the US Army Research Laboratory (Rome, New York). He has served as Chair of the IEEE Signal rocessing Society Technical Committee on Design and Implementation of Signal Processing Systems. He is a Fellow of the IEE.

We are surrounded by millions of cameras in our everyday life. However, traditional camera networks require expensive infrastructures and are typically not able to provide user-oriented services. Smart cameras may overcome these limitations; they perform substantial image processing onboard delivering only features of the observed scene and collaborate to overcome some problems of centralized or single-camera systems. This interdisciplinary field builds upon techniques from computer vision, distributed computing, embedded computing and sensor networks. Pervasive smart cameras integrate adaptivity and autonomy and support a service-oriented network which is easy to deploy and operate, adapts to changes in the environment and provides various customized services to users.
In this talk I will introduce smart cameras and their potential for various applications such as smart environments, security, entertainment and health care. I will then focus on the fundamental challenges of performing real-time vision on distributed embedded platforms and address recent research topics. A presentation of case studies of distributed smart cameras will conclude this talk.

Visuelle Sensornetze erfassen, verarbeiten und übertragen Bilddaten in Echtzeit und finden in unterschiedlichen Bereichen, wie beispielsweise Automotive, Security und Pervasive Computing, verstärkt Anwendung.  Diese verteilten, eingebetteten Systeme stellen besondere Herausforderungen an die Nutzung der beschränkt verfügbaren Ressourcen, wie Rechenkapazität, Speicher, Übertragungskapazität und Energie, dar. In diesem Vortrag werde ich diese Herausforderungen kurz erläutern und verschiedene Ansätze für den ressourcen-effizienten Entwurf und Betrieb von visuellen Sensornetzen auf Knoten- und Netzwerkebene vorstellen sowie ausgewählte Anwendungen demonstrieren.

Smart cameras combine video sensing, processing, and communication on a single embedded platform. Networks of smart cameras are real-time distributed embedded systems that perform computer vision using multiple cameras. This new approach has emerged thanks to a confluence of simultaneous advances in four key disciplines: computer vision, image sensors, embedded computing, and sensor networks. Recently these visual sensor networks have gained a lot of interest in research and industry; applications include surveillance, assisted living and smart environments. This tutorial focuses on the networking aspects of smart camera systems where visual data is processed in real-time using distributed sensing and computing nodes. Although this distribution of sensing and processing introduces several complications, we believe that the problems it solves are much more important than the challenges of designing and building a distributed smart camera network. As in many other applications, distributed systems scale much more effectively, require less network bandwidth and achieve shorter response times than do centralized systems. We conclude this tutorial by a description of applications and case studies of visual sensor networks. We further discuss recent trends of this exciting research field.

Distributed smart cameras are real-time distributed embedded systems that perform computer vision using multiple cameras. Smart cameras perform substantial image processing onboard delivering only features of the observed scene and collaborate to overcome some problems of centralized or single-camera systems. This interdisciplinary field builds upon techniques from computer vision, distributed computing, embedded computing and sensor networks. In this talk I will introduce smart cameras and their potential for various applications such as smart environments, security, entertainment and health care. I will then focus on the fundamental challenges of performing real-time vision on distributed embedded platforms and address recent research topics. A presentation of case studies of distributed smart cameras will conclude this talk.

Recently, computer vision has gradually been making the transition away from understanding single images to analyzing image sequences, or video understanding. Video understanding deals with understanding video sequences, e.g., recognition of gestures, activities, and facial expressions. The main shift in the classic paradigm has been from the recognition of static objects in the scene to motion-based recognition of actions and events. Since most videos are about people, this work has mainly focused on analysis of human motion. In particular, there has been a significant interest in the automated visual surveillance systems. Such systems have the advantage of providing continuous active warning capabilities and are especially useful in the areas of law enforcement, national defense, border control, and airport security. The main steps in video understanding are: detection of objects of interest in video (e.g. people, vehicles), tracking of those objects from frame to frame, and recognition of their activities and behavior. In this talk, I will present an overview of our work in video understanding.

Distributed smart cameras are real-time distributed embedded systems that perform computer vision using multiple cameras.  Smart cameras perform substantial image processing onboard delivering only features of the observed scene and collaborate to overcome some problems of centralized or single-camera systems.  This new approach is emerging thanks to a confluence of demanding applications and the huge computational and communications abilities enabled by Moore’s Law.  This interdisciplinary field builds upon techniques from computer vision, distributed computing, embedded computing and sensor networks. 

Distributed smart cameras are real-time distributed embedded systems that perform computer vision using multiple cameras.  Smart cameras perform substantial image processing onboard delivering only feature

s of the observed scene and collaborate to overcome some problems of centralized or single-camera systems.  This new approach is emerging thanks to a confluence of demanding applications and the huge computational and communications abilities enabled by Moore’s Law.  This interdisciplinary field builds upon techniques from computer vision, distributed computing, embedded computing and sensor networks.

Seit vielen Jahren beobachten wir eine massive Verbreitung von Kamerasystemen im öffentlichen, beruflichen und privaten Umfeld. Getrieben vom rasanten technologischen Fortschritt, dem damit einher gehenden Preisverfall aber auch von einem vermeintlich gesteigerten Sicherheitsbewusstsein hat sich dieser Trend in den letzten Jahren beschleunigt. Wir sind zwar nun von vielen (Kamera-)Augen umgeben, aber ihre „Sehfähigkeit“ – im Sinne von Wahrnehmung und Erkenntnis – ist noch sehr beschränkt.
Kamerasysteme werden nicht nur zur Überwachung von Gebäuden, Verkehrswegen und ähnlichem eingesetzt, man findet sie beispielsweise auch in der industriellen Fertigung, der Robotik, der Materialprüfung und im Mobiltelefon. Eine Speicherung bzw. manuelle Analyse aller aufgenommenen Videoströme scheitert unter anderem an der schieren Datenmenge. Hier sind innovative Architekturen und Bildverarbeitungsmethoden für die Kamerasysteme gefordert. „Intelligente Kameras“ bieten einen Ansatz zur Lösung dieser Probleme. Sie vereinen Bildaufnahme, Bildverarbeitung und Kommunikation der analysierten Videodaten in einem eingebetteten System. Intelligente Kameras liefern keine Bilder, sondern die Ergebnisse der Bildanalyse, wie z.B. die Koordinaten eines bewegten Objektes oder die Kennzeichen von vorbeifahrenden Fahrzeugen. Sie arbeiten gemeinsam in einem Kamera-Netzwerk und erhöhen dadurch ihre „Sehfähigkeit“. Intelligente Kameras repräsentieren ein sehr aktives Forschungsgebiet im Schnittpunkt der Bereiche Bildverarbeitung, Sensor Netzwerke, eingebettete und verteilte Systeme sowie Pervasive Computing. In diesem Vortrag werde ich aktuelle Kamerasysteme kurz vorstellen und ihre Grenzen aufzeigen. Anschließend möchte ich auf „intelligente Kameras“ sowie ihre zugrunde liegenden Methoden näher eingehen. Fallstudien und Anwendung dieser Kameras im Gesundheitswesen, im Unterhaltungsbereich bzw. in der Verkehrsüberwachung bilden den Abschluss des Vortrages.

Tutorial on Distributed Vision Processing in Smart Camera Networks

Co-Lecturer: Hamid Aghajan (University of Stanford), Francois Berry (Clermont University), Horst Bischof (Graz University of Technology), Richard Kleihorst (NXP Research), Wayne Wolf (Princeton University)

Title: Embedded Middleware on Distributed Smart Cameras (invited paper)