Day One

Computer Vision Methods

10:00 - 10:30 - Repeatability Experiments for Interest Point Location and Orientation Assignment

Simon Taylor (MIL, Engineering Department / Cambridge University)

The initial stage in many keypoint-based matching schemes is to locate interest points and assign them some canonical parameters, such as orientation and scale, in a viewpoint-invariant manner. The repeatability of this first stage is a key factor in overall matching performance. I present the results of some experiments on the repeatability of various interest point detectors and orientation assignment methods. A number of orientation assignment schemes for FAST features are proposed, and the FAST-specific methods are compared to other common orientation assignment schemes based on image gradients.

10:30 - 11:00 - Real-time interactive 3D reconstruction

Qi Pan (MIL, Engineering Department / Cambridge University)

All high quality 3D reconstruction algorithms currently require an input sequence of images or video which is then processed offline for a lengthy amount of time. After the process is complete, the reconstruction is viewed by the user to confirm the algorithm has modelled the input sequence successfully. Often it is the case that certain parts of the reconstructed model may be inaccurate or sections may be missing due to insufficient coverage or occlusion in the input sequence. In these cases, a new input sequence needs to be obtained and the whole process repeated. The aim of the project I am working on is to produce a real-time modelling system which provides immediate feedback about the quality of the input sequence. This enables the system to guide the user to provide additional views for reconstruction, yielding a complete model without having to collect a new input sequence.

11:00 - 11:30 - Computer Vision and Augmented Reality

Vincent Lepetit (Computer Vision Lab / EPFL)

I will present recent advances in Computer Vision done at CVLab for Augmented Reality. I will also discuss the limits of the current techniques, and the possible directions to go beyond.

11:30 - 11:45 - Coffee Break

Keynote Speech

11:45 - 13:00 - Projection Over Four Orders of Magnitude

Keynote by Oliver Bimber (Faculty of Media / Bauhaus University Weimar)

Lunch/Lab Visit

13:00 - 14:00 - Lunch (at Mensa)

14:00 - 14:45 - Lab Visits

Deskoteque, HEyeWall, DAVE, VisCenter

Applications

14:45 - 15:15 - VirtualStudio2Go: Digital Video Composition for Real Environments

Anselm Grundhöfer (Faculty of Media / Bauhaus University Weimar)

We synchronize film cameras and LED lighting with off-the-shelf video projectors. Radiometric compensation allows displaying keying patterns and other spatial codes on arbitrary real world surfaces. A fast temporal multiplexing of coded projection and flash illumination enables professional keying, environment matting, displaying moderator information, scene reconstruction, and camera tracking for non-studio film sets without being limited to the constraints of a virtual studio. The reconstruction of the scene geometry allows special composition effects, such as shadow casts, occlusions and reflections. This makes digital video composition more flexible, since static studio equipment, such as blue screens, teleprompters, or tracking devices, is not required. Authentic film locations can be supported with our portable system without causing a lot of installation effort. We propose a concept that combines all of these techniques into one single compact system that is fully compatible with common digital video composition pipelines, and offers an immediate plug-and-play applicability.

15:15 - 15:45 - Mobile Augmented Reality Showcase: Wikitude AR Travel Guide

Philipp Breuss-Schneeweis (Mobilizy Project / mobilizy.com)

Wikitude is a mobile travel guide for the Android platform based on location-based Wikipedia content. It is a handy application for planning a trip or to find out about landmarks in your surroundings; 350,000 world-wide points of interest may be searched by GPS or by address and displayed in a list view, map view or cam view.

15:45 - 16:15 - HYDROSYS

Ernst Kruijff (ICG / Graz University of Technology)

HYDROSYS deals with on-site monitoring of environmental resources, using (among others) handheld AR and robotics. This talk will introduce the project's approach and technical framework, lining out some of its innovative aspects.

16:15 - 16:30 - Break

Tracking I

16:30 - 17:00 - Rigid-Body Target Optimization for Optical Motion-Tracking

Thomas Pintaric (IMS / Vienna University of Technology)

We propose a simple, yet effective new methodology for constructing optically tracked rigid-body targets. Our approach employs an optimization heuristic to compute near-optimal marker arrangements for an arbitrary number of targets with the goal to minimize marker labeling errors at runtime. New rigid-body targets can be optimized for joint use with an existing set of targets. Using prefabricated mounting fixtures, the target assembly step requires only basic hobbyist tools and skills.

17:00 - 17:30 - Finger-Detection for Search Space Decomposition in Image-Based Hand Posture Tracking

Nils Petersen (DFKI)

There are two main challenges in hand posture tracking. The first one is the high dimensionality of the configuration space of about 27 DOFs. The second one is, that the hand and fingers can be moved at high velocity, leading to highly dissimilar consecutive images at standard camera frame rates around 30 fps. The use of a motion model of the hand is only of modest utility, since accelerations are likely to vary too quickly as well. Our method makes the reasonable assumption that there is no occlusion by the palm or foreground objects. In that case there must always be at least one finger fully visible in the image. We then start by detecting fingers with a simpler 2D-model first. The collected information is reformulated into hypotheses that are used to decompose the actual search space by fixing the DOFs for the given finger, in turn assuming it being the thumb, the index, etc. For an efficient implementation we investigate the use of particle filters within a Dependent Bayesian Filter Integration (DBFI) Framework.

17:30 End

Evening

20:00 Dinner + Social Event

Location to be announced

Day Two

9:30 - 9:45 - News on ISMAR 2009

Christine Perey (PEREY Research & Consulting )

Tracking II

9:45 - 10:15 - Bluetooth Tracking without Discoverability

Simon Hay (DTG / University of Cambridge)

Outdoor location-based services are now prevalent due to advances in mobile technology and GPS. Indoors, however, even coarse location remains unavailable. Bluetooth has been identified as a potential location technology that mobile consumer devices already support,easing deployment and maintenance. However, Bluetooth tracking systems to date have relied on the Bluetooth inquiry mode to constantly scan for devices. This process is very slow and can be a security and privacy risk. In this paper we investigate an alternative: connection-based tracking. This permits tracking of a previously identified handset within a field of fixed base stations. Proximity is determined by creating and monitoring low-level Bluetooth connections that do not require authorisation. We investigate the properties of the low-level connections both theoretically and in practice, and show how to construct a building-wide tracking system based on this technique. We conclude that the technique is a viable alternative to inquiry-based Bluetooth tracking.

10:15 - 10:45 - Real-time and robust natural feature tracking on mobile phones

Daniel Wagner (ICG / Graz University of Technology)

We present a new natural feature tracker that combines two different methods for detection and tracking into a new hybrid tracking system at previously unseen performance and robustness on mobile phones. This new system is faster than previous natural feature trackers, tracks under extreme tilts, is robust to blur and works over large scale changes.

10:45 - 11:00 - Break

User-centric

11:00 -11:30 - Simulating Educational Physical Experiments in Augmented Reality

Hannes Kaufmann (IMS / Vienna University of Technology)

We present an augmented reality application for mechanics education. It utilizes a recent physics engine developed for the PC gaming market to simulate physical experiments in the domain of mechanics in real time. Students are enabled to actively build own experiments and study them in a three-dimensional virtual world. A variety of tools are provided to analyze forces, mass, paths and other properties of objects before, during and after experiments. Innovative teaching content is presented that exploits the strengths of our immersive virtual environment. PhysicsPlayground serves as an example of how current technologies can be combined to deliver a new quality in physics education.

11:30 - 12:00 - Map Torchlight: A Mobile Augmented Reality Camera Projector Unit

Johannes Schöning (University of Münster)

The advantages of paper-based maps have been utilized in the field of mobile Augmented Reality (AR) in the last few years. Traditional paper-based maps provide high-resolution, large-scale information with zero power consumption. There are numerous implementations of magic lens interfaces that combine high-resolution paper maps with dynamic handheld displays. From an HCI perspective, the main challenge of magic lens interfaces is that users have to switch their attention between the magic lens and the information in the background. In this paper, we attempt to overcome this problem by using a lightweight mobile camera projector unit to augment the paper map directly with additional information. The "Map Torchlight" is tracked over a paper map and can precisely highlight points of interest, streets, and areas to give directions or other guidance for interacting with the map.

12:00 - 12:30 - Saliency Modulation for Focus+Context from Video

Erick Mendez (ICG / Graz University of Technology)

In this talk we present a technique for automatic video based Focus and Context (F+C) discrimination for Augmented Reality. Traditionally, selection of the F+C strategy is either statically or manually given by the user. This restriction is owed to the dependency of the strategies to the objects currently visible in the scene. In contrast, our technique is based on an analysis of the current video feed. We inspect the image saliency map, and modify the image to favor our focus object. Previous work in perceptual psychology shows evidence that there is a correspondence between the image’s saliency and the user’s attention. We exploit this correspondence to draw the user’s attention to our desired focus object by emphasizing its saliency.

Lunch and Demos

12:30 - 13:30 - Lunch (at Mensa)

13:30 - 14:00- Demos

Systems

14:00 - 14:30 - Spatial Tags: Distributed Spatial Indexing using Fiducial Markers

Tom Craig (DTG / University of Cambridge)

Fiducial markers, or tags, have long been used in the augmented reality community to enable a machine vision system to determine the position and orientation of an observer. We have extended these tags to represent a general spatial zone which, by associating the tag directly with a computational resource, can be used to trigger an action upon ingress or egress of the zone; we call such markers "Spatial Tags". Our system has many desirable properties: it is decentralised, each zone is trivial to create or modify, and behaviour is intuitive to its users. We evaluate these advantages by using our Spatial Tags in a "follow-me desktop" application.

14:30 - 15:00 - The HEyeWall - a 4m x 2m multitouch screen

Marcel Lancelle (CGV / Graz University of Technology)

With 8 Megapixels the HEyeWall is a large high resolution rear projection wall. Our novel projection technology eliminates seams by using a large low resolution projector plus six small high resolution projectors for the details. The FTIR multitouch functionality will cover the whole screen and is still work in progress.

15:00 - 15:30 - Autonomous explorations in a smart room

Peter Holzer (VMG / Graz University of Technology)

Autonomous systems have been subject to research for years. Such systems are able to interact with the real word without user control. Thus, the human-in-the loop is replaced by a machine-in-the-loop. This paper presents a highly modular autonomous exploration system, with the purpose to automatically navigate an autonomous agent close-by to a user selected object. The system as such consists of a movable core system and autonomous subsystems. In particular, the core system is based on a dual-camera structure and motion framework. The autonomous subsystem is presented by a computer, connected to a remotely controlled mobile platform (RC-car) via a custom hardware interface. Additionally, subsystems with pan-tilt zoom cameras on tripods may be used. The RC-car itself features a single camera on a pan-tilt unit, which transmits data to the computer via radio link. The core- and subsystem are interconnected via a wireless connection. The RC-car is tagged by a marker, to relieve the problem of pose retrieval of the autonomous agent. In contrast to permanent marker tracking, the core system requires the marker only once; the trajectory as well as the 3D scene structure is computed from natural landmarks. By manually selecting a certain object in the scene, the core system is capable of instructing the autonomous RC-car to navigate towards the object. However, these instructions are only required at initialization, the subsystem is able to navigate autonomously.

15:30 - 15:45 - Break

Reports from Practice

15:45 - 16:15 - Taking Mixed and Augmented Reality to the Mass Market

Christine Perey (PEREY Research & Consulting )

Moving from research to commercial success is a messy affair. There are many different motivations or incentives. And there are as many strategies as there are experts. Will Mixed and Augmented Reality ever achieve mass market penetration? What are appropriate stages of evolution we can expect in the maturation of mixed and augmented reality technologies? How will you choose the best partner? What role does your research or project play in the forming of successful strategies and your ability to achieve your or your go-to-market partner's objectives? This session is designed to provoke thought and discussion about the future of mixed and augmented reality and to provide concrete suggestions for investment in the present.

16:15 - 16:45 - Commercial Mobile Augmented Reality Development: Where Does All the Research Money End Up?

Istvan Barakonyi (Imagination)

The development and deployment of commercial AR applications requires active cooperation, communication, and understanding between researchers, coders, graphic designers, and most importantly customers. This talk will give an insight into how this collaboration works at Magitech, a Vienna-based company, and presents case studies of successfully deployed commercial mobile AR applications including their development and authoring pipeline.

16:45 - 17:15 - How to get the next generation hooked on augmented reality Today

Ori Inbar (Pookatak Games)

Our belief: in 10-15 years everyone will use Augmented Reality to experience the world in a more meaningful way. Our collective mission: to nurture a healthy industry that will drive the adoption sooner than later. So where do we start? I argue that we should start with the younger generation, a generation that is ripe and eager to try new experiences that speak their language. The challenge is how to give them something they enjoy, and at the same time offers value to those who hold the buying power - their parents, guardians, or teachers. In this talk, I will describe the philosophy, the challenges, and the resulting design and strategy behind an educational handheld AR game - currently in development - and scheduled to hit consumers in the first half of 2009.

17:15 Finishing Words

Evening

20:00 Dinner

Location to be announced

Back to main page