Applying I-FGM to image retrieval and an I-FGM system performance analyses

International Journal of Electronic Government Research

Nguyen³

et al. 2008

Self Cite

Homeland securiEy and disaster reliefore some oflhe crilical Uri'US o/t.:-gol'ernance Ihat have to deal with vast amounts ofdynamic heterogeneous Jala. Providing rapid reul-lime search capobilitiesjor such (Ippli calions is a challenge. Intelligent Foraging, Galhering, and .\latching (I-FGM) is an estublishedji"Wllell'ork developed to assistllsers 10 find infornwlion qllickl)' and eJfective(v by incremelJ/all)' collecting. lil'Ocessing and malching information nllggels This jraTilell'Ork has been sllccessfitlly IIsed to develop a distribll/ed, unstructured lexl retrieval application. In this papeJ: we apply the I-FGA1 framework 10 image collections by using a concept-based image retrieval method H"e {lpproaclz fhis by incrementally processing images. ex tracting low-level jeatures and mapp;ng them 10 higher level concepts. Our empiricol evaluation sholl's that 0111' approach performs competitively compared 10 some existing approaches in terms ofretrieving relevant images while offering Ihe speed advalJ/oges o/distrlbliled and incremental piVcess and unifiedjramclI'ork between lexl and images

Section: Discussion Results and Analysismentioning

confidence: 99%

Section: Validation Of Image Retrieval Algorithmmentioning

confidence: 95%

I-FGM as a Real Time Information Retrieval Tool for E-Governance

International Journal of Electronic Government Research

Nguyen³

et al. 2008

Self Cite

“…HM1582-04-1-2027. Some preliminary results in this paper can be found in [28,29]. Also, thanks to Morgan Pittkin for his assistance in this project.…”

Section: Acknowledgementsmentioning

confidence: 95%

“…In this paper, we present such a large-scale distributed real-time framework, called the anytime Intelligent Foraging, Gathering, and Matching (I-FGM) system [27][28][29]31]. We integrate the knowledge representation techniques from Artificial Intelligence (AI) and algorithms from Distributed and Parallel computing to efficiently allocate and deploy computation resources in a unified framework.…”

Section: Introductionmentioning

confidence: 99%

A large-scale distributed framework for information retrieval in large dynamic search spaces

Nguyen

et al. 2010

Appl Intell

Self Cite

One of the main problems facing human analysts dealing with large amounts of dynamic data is that important information may not be assessed in time to aid the decision making process. We present a novel distributed processing framework called Intelligent Foraging, Gathering and Matching (I-FGM) that addresses this problem by concentrating on resource allocation and adapting to computational needs in real-time. It serves as an umbrella framework in which the various tools and techniques available in information retrieval can be used effectively and efficiently. We implement a prototype of I-FGM and validate it through both empirical studies and theoretical performance analysis.

“…The WALRUS-based I-FGM prototype was the first step towards deploying image retrieval in I-FGM (Santos et al, 2007). It was developed to show that a regions based image Here P and Q represent the retrieved image and the query image respectively.…”

Section: Validation Of Image Retrieval Algorithmmentioning

confidence: 99%

Analyst-Ready Large Scale Real Time Information Retrieval Tool for E-Governance

E-Government Diffusion, Policy, and Impact

Nguyen

et al. 2009

Self Cite

With the proliferation of the Internet and rapid development of information and communication infrastructure, E-governance has become a viable option for effective deployment of government services and programs. Areas of E-governance such as Homeland security and disaster relief have to deal with vast amounts of dynamic heterogeneous data. Providing rapid real-time search capabilities for such databases/sources is a challenge. Intelligent Foraging, Gathering, and Matching (I-FGM) is an established framework developed to assist analysts to find information quickly and effectively by incrementally collecting, processing and matching information nuggets. This framework has previously been used to develop a distributed, free text information retrieval application. In this chapter, we provide a comprehensive solution for the E-GOV analyst by extending the I-FGM framework to image collections and creating a “live” version of I-FGM deployable for real-world use. We present a Content Based Image Retrieval (CBIR) technique that incrementally processes the images, extracts low-level features and map them to higher level concepts. Our empirical evaluation of the algorithm shows that our approach performs competitively compared to some existing approaches in terms of retrieving relevant images while offering the speed advantages of a distributed and incremental process, and unified framework for both text and images. We describe our production level prototype that has a sophisticated user interface which can also deal with multiple queries from multiple users. The interface provides real-time updating of the search results and provides “under the hood” details of I-FGM processes as the queries are being processed.