Showing papers by "Peter Bajcsy published in 2008"

Integration of thermal and visible imagery for robust foreground detection in tele-immersive spaces

Abstract: Tele-immersive systems are designed to facilitate communication and collaboration between people. Therefore, the central objects of interest in a tele-immersive system are these people, the things they jointly manipulate, and the tools they need to perform this manipulation. This paper presents a multi-modal image fusion framework to detect these objects of interest. Experimental results with a prototype tele-immersive system show that fusion of visible and thermal imagery enables robust foreground detection in unstructured environments. The contribution of this work lies in designing the integration framework, prototyping hardware and software components, and evaluating quantitatively the multi-modal foreground detection for a set of standard scenarios.

Automated Placement of Multiple Stereo Cameras

Abstract: This paper presents a simulation framework for multiple stereo camera placement. Multiple stereo camera systems are becoming increasingly popular these days. Applications of multiple stereo camera systems such as tele-immersive systems enable cloning of dynamic scenes in real-time and delivering 3D information from multiple geographic locations to everyone for viewing it in virtual (immersive) 3D spaces. In order to make such multi stereo camera systems ubiquitous, solving the problem of optimal deployment (configuration) of 3D imaging components is motivated by the need (a) to create high quality 3D content and (b) to accommodate application specific requirements into optimal deployment without ad-hoc experimentations. One of the configuration parameters is the placement of stereo cameras that a®ects the quality of 3D reconstructions as well as the resolution achieved by the reconstruction. The novelty of our work is in formulating an optimization framework for optimal camera placement using error based objective function and five constraints. We generate an initial solution to the optimization problem using Genetic Algorithms and then refine the solution using Gradient Descent. The algorithm is validated using actual camera placement as well as using simulation results. The results not only show promising features of our optimization approach but also eliminate ad-hoc experimentation of camera placement for each end application where multiple stereo camera systems can be deployed.

Text, Image and Vector Graphics Based Appraisal of Contemporary Documents

Abstract: We have designed a framework for content based appraisal of documents. Our motivation is to provide computer assisted support for answering several appraisal criteria according to the general appraisal guidelines in the National Archives and Record Administration (NARA) 1441 directive. The appraisal criteria led us to investigations related to (a) finding groups of PDF documents with similar content, (b) ranking documents according to their creation/ modification time and digital volume, and (c) detecting inconsistency between ranking and content within a group of related documents. The novelty of our work is in designing a methodology and a mathematical framework for document appraisals, and prototyping the framework working with text, image and vector graphics components of PDF documents. We present example results of grouping, ranking and integrity verification for groups of scientific documents about medical topics.

A Perspective on Cyberinfrastructure for Water Research Driven by Informatics Methodologies

Abstract: This article presents a perspective on cyberinfrastructure (CI) for water research driven by informatics methodologies. This perspective is motivated by the fact that CI for water research should increase scientific productivity of a single investigator and dispersed teams in their exploratory studies of experimental observations and theoretical simulations. In the digital era and within the context of CI, the scientific productivity is often determined by the efficiency of data-centric and collaborationcentric activities. These activities follow domain-specific methodologies and informatics approaches aiming at extracting information and knowledge from raw data. We present the concepts of data-centric and collaboration-centric activities supported by concrete examples, and outline the challenges and requirements on CI driven by these informatics activities. Then, we describe a set of common activities cutting across multiple earth science disciplines, and discuss some solutions that already exist or might have to be developed in order to support informatics methodologies. Our perspective emphasizes the importance of exploratory science that is frequently present in informatics methodologies. The contribution of this article is in illuminating the CI development from a perspective that bridges daily activities of many water research scientists with the CI components and functionality.

Publishing Active Workflows to Problem-Focused Web Spaces

Abstract: We present a new paradigm for publishing scientific workflows to the Web. Publishing in this system involves not only creating a Web space for running the workflow on demand, but also the ability for the publisher to convey expertise about its use by providing ways to focus attention on problem-relevant parameters, inputs and outputs. As part of the publishing step we include the ability to configure visualizations for specific outputs of the workflow. The resulting system is a Web based environment focused on the ability to publish existing workflows so that other users in the community can execute those workflows and visualize the results from the Web. Workflow executions and visualizations can be shared with the community along with workflow definitions.

Real time localization and 3D depth estimation across disparate sensing systems: Toward hazard aware and tele-immersive spaces

Abstract: This paper addresses the problems of real-time localization and 3D depth estimation across disparate sensing systems. The sensing systems include wireless microelectromechanical systems (MEMS) sensor networks, such as MICA sensors by Crossbow Inc., radio frequency identification (RFID) tags and cameras that capture a variety of spectra. Some of the sensing is adaptive in time and space by using a remotely controlled robot for the sensor deployment. The motivation for integrating and analyzing multiple sensing systems and spectral modalities comes from the fact that in many applications a single sensing system or modality does not lead to robust and accurate performance. In this work we design systems for localization using radio frequency identification (RFID) tags and real time 3D depth estimation from stereo vision in order to incorporate power constraints imposed on deployment of battery-operated wireless MICA sensors. The resulting methods are applied to the development of (a) hazard aware spaces (HAS) to alert people in events of fire, and (b) tele-immersive spaces (TEEVE) to enable remote collaborations, training and art performances. The novelty of our work lies in the power efficient deployment of wireless sensors for location aware applications by combining multiple sensors with advanced signal and image processing algorithms.

Knowledge Spaces and Scientific Data

Abstract: As scientific work becomes more complex, contextual information is critical to enable its use. We have developed the concept of Knowledge Spaces that manage such information and scope the space of data, metadata, and algorithms relevant to the work of a Virtual Organization. This enables scientists to produce and manage descriptive metadata, process description, data provenance, as well as VO-specific quality assertions, vocabulary mappings, inferences, computation, approximations, and security policies. We have developed middleware and cooperating applications that demonstrate the potential of Knowledge Spaces in the production and dissemination of scientific data.

Advanced Information Delivery System for the Abraham Lincoln Writings

Abstract: This work presents a virtual observatory of the Abraham Lincoln writings. Many of the Lincoln correspondence papers and legal documents have been digitized, pre-processed and annotated for public viewing. However, there are challenges in delivering the information to the general public due to the large number of documents, very large volume of digital files, lack of metadata, heterogeneous pieces of information and need for services such as searching and transcription. We have prototyped an advanced information system that delivers Lincoln writings together with historical maps, contemporary Google maps, and Lincoln Log metadata, and other types of information to provide a multi-dimensional view of Lincoln's life. The novelty of our work is in designing the system that provides such a unique multi-dimensional view of Lincoln's life by combining technologies needed for building virtual observatories.

Advanced Information Systems for Archival Appraisals of Contemporary Documents

Abstract: This work addresses the problem of designing a scalable framework for archival appraisals of contemporary PDF documents. The motivation for our work is to provide an e-science solution that (a) fuses the independent research methodologies focusing on specific information types to one comprehensive analytical framework, (b) optimizes tradeoffs between computational requirements and preservation costs, and (b) bridges the small scale and large scale computational studies. The e-science solution presented here consists of (1) a methodology for comprehensive comparisons of contemporary documents containing text, images and vector graphics, (2) a framework for including 3D and 3D+time data sets into the appraisal analyses, (3) interfaces supporting exploratory archival appraisal analyses with small scale data sets, and (4) infrastructure supporting the transition from small scale to large scale computations using commodity and high performance computing resources. The novelty of our work is in designing methodologies, mathematical frameworks and prototypes for comprehensive and scalable document appraisals that include text, images, vector graphics, and high dimensional data.