Dr. Shu-Ching Chen is a Professor in the School of Computing and Information Sciences (SCIS) at Florida International University (FIU), Miami. He has been a Full Professor since August 2009 in SCIS at FIU. Prior to that, he was an Assistant/Associate Professor in SCIS at FIU from 1999. He received his Ph.D. degree in Electrical and Computer Engineering in 1998, and Master's degrees in Computer Science, Electrical Engineering, and Civil Engineering in 1992, 1995, and 1996, respectively, all from Purdue University, West Lafayette, IN, USA.
He is the Director of the Distributed Multimedia Information Systems Laboratory (DMIS) and the Co-Director of the Integrated Computer Augmented Virtual Environment (I-CAVE). His main research interests include multimedia big data, content-based image/video retrieval, multimedia data mining, multimedia systems, and Disaster Information Management. As the Director of the DMIS lab, Dr. Chen has led the development of challenging, innovative, and community-centered projects, among which stands out the Florida Public Hurricane Loss Model (FPHLM), of which he is the Co-PI and leader of the Computer Science Component.
Dr. Chen has authored and coauthored more than 350 research papers in journals, refereed conference/symposium/workshop proceedings, book chapters, and four books. Dr. Chen has been the PI/Co-PI of many research grants from NSF, National Oceanic and Atmospheric Administration (NOAA), Department of Homeland Security, Army Research Office, Naval Research Laboratory (NRL), Florida Office of Insurance Regulation, IBM, and Florida Department of Transportation with a total amount of more than 35 millions.
Dr. Chen was named a 2011 recipient of the ACM Distinguished Scientist Award. He received the best paper awards from 2006 IEEE International Symposium on Multimedia and 2016 IEEE International Conference on Information Reuse and Integration. He was awarded the IEEE Systems, Man, and Cybernetics (SMC) Society's Outstanding Contribution Award in 2005 and was the co-recipient of the IEEE Most Active SMC Technical Committee Award in 2006. He was also awarded the Top Scholar Award from FIU in 2011 and 2012, Inaugural Excellence in Graduate Mentorship Award from FIU in 2006, the University Outstanding Faculty Research Award from FIU in 2004, the Excellence in Mentorship Award from SCIS in 2010, the Outstanding Faculty Service Award from SCIS in 2004 and 2014, and the Outstanding Faculty Research Award from SCIS in 2002, 2012, and 2017, and Director's Special Recognition Award from SCIS in 2016. He is a fellow of IEEE, AAAS, and SIRI.
He has been a General Chair and Program Chair for more than 60 conferences, symposiums, and workshops. He is the Editor-in-Chief of IEEE Multimedia Magazine, founding Editor-in-Chief of International Journal of Multimedia Data Engineering and Management, and Associate Editor/ Editorial Board of IEEE Trans. on Multimedia and IEEE Trans. on Human-Machine Systems. He served as the Chair of IEEE Computer Society Technical Committee on Multimedia Computing (TCMC) from 2010 to 2015. TCMC is one of the 18 technical committees/councils who received an excellent vitality rating in 2013. IEEE Computer magazine ran a piece on the 2013 technical committee vitality review process in the December 2013 issue. Under Dr. Chen's leadership, the TCMC received excellent reviews during five consecutive years (2011-2015). He is also Co-Chair of IEEE Systems, Man, and Cybernetics Society's Technical Committee on Knowledge Acquisition in Intelligent Systems. Dr. Chen also has been a guest editor for more than ten journal special issues. He was a member of three steering committees (including IEEE Transactions on Multimedia) and several panels for conferences and NSF. He also serves/served as a member of technical program committee for more than 360 professional meetings. Dr. Chen is also the Co-Founder of Bay Area Multimedia Forum.
As the Director of the Distributed Multimedia Information Systems Laboratory at SCIS, Dr. Chen has worked with undergraduate and graduate students on the design, implementation, and maintenance of challenging, far-reaching, and community-centered projects.
Dr. Chen is the Co-PI and leader of the Computer Science Component of the Florida Public Hurricane Loss Model (FPHLM), a large-scale hurricane catastrophe model developed by a multi-disciplinary team of experts in the fields of meteorology, wind and structural engineering, computer science, GIS, statistics, finance, and actuarial science. It is the first public hurricane loss model in the world. The realization of this innovative project involves the close collaboration of multiple academic and scientific institutions, of which Florida International University is the leading institution.
The state of Florida ranks number one in total insured property value exposed to hurricane wind. Out of Florida's $3.6 trillion in insured properties, about $2 trillion are residential, and all are exposed to hurricane risk. This reality gave birth to this project. Funded by the Florida Office of Insurance Regulation (FL OIR), the FPHLM was born out of the need of accurately assessing hurricane wind risk and predicting insured losses for the Floridians' residential properties. Since the release of its first version in March of 2006 (see press release), the model has been used over 2,600 times by the state and over 200 times by firms in the insurance industry. As an open project, which undergoes a strict certification process every two years, the FPHLM provides the state regulators a reliable benchmark in the homeowner insurance market and allows them to better formulate the state's homeowner insurance rate-making policy. The FPHLM's impact on the community is unquestionable, as its predictions directly influence the home insurance rates payed by Floridians every month.
The model consists of three major components: wind hazard (meteorology), vulnerability (engineering), and insured loss cost (actuarial). It has over a dozen sub-components. The computer platform is designed to accommodate future hookups of additional sub-components or enhancements.
The FPHLM estimates loss costs and probable maximum loss levels from hurricane events for personal lines and commercial lines of residential property. The losses are estimated for building, appurtenant structures, contents, and additional living expenses.
In essence, the FPHLM is a complex collection of computer programs that simulate and predict how, where and when hurricanes form, their wind speeds, intensity and sizes, their tracks, how they are affected by the terrain after landfall, how the winds interact with different types of structures, how much damage they can cause to house roofs, windows, doors, and interiors, how much it will cost to rebuild the damaged parts, and how much of the loss will be paid by insurers.
Florida also ranks number one in coastal property exposed to storm surge. In the face of this reality, in 2013 the state funded FIU to enhance the FPHLM by adding both a storm surge component and an inland flooding component. The proposed new model will assess storm surge and hurricane-related rain flood risk and estimate both the insured and uninsured losses that they may cause. The enhancement project will take three years with $4.5 million funding from the State.
More information on the FPHLM is available on the project's web site.
Miami Herald - Florida property insurers get high marks in hurricane 'stress test' - We ran the FPHLM model for FL-OIR in Summer 2015.
Miami-Dade County is particularly susceptible to storm surges caused by hurricanes. To help Miami-Dade county residents understand the dangers of storm surge, hydrological and software development experts at FIU's International Hurricane Research Center (IHRC) and College of Engineering and Computing (CEC) teamed up with Miami-Dade County to create a Storm Surge Simulator, a web-based application that allows users to visualize the potential storm surge effects on their home or business.
Dr. Chen led the development team of Computer Science students to combine cutting-edge web technologies with storm surge modeling output in order to help the Miami-Dade community.
The result of the paper A Progressive Morphological Filter for Removing Non-Ground Measurements from Airborne LIDAR Data, authored by Dr. Chen and his colleagues K. Zhang, D. Whitman, M.-L. Shyu, J. Yan, and C. Zhang, and published in IEEE Transactions on Geoscience and Remote Sensing, vol. 41, issue 4, pp. 872-882, April 2003, was used by the official Miami-Dade County Storm Surge Simulator project. This paper has been cited 988 times by Google Scholar.
The Storm Surge Simulator has been reported by many media outlets including the Miami Herald, the Florida Statesman, Mapperz - The Mapping News Blog, the AnyGeo Blog, the MapThis! Blog, Google Maps Mania, WIOD 610AM, Diario Las Americas, NBC 6, CBS 4, the Emergency Management Magazine, and the Government Technology Magazine.
The Storm Surge Simulator is available to the public at no charge. Click here to visit the Storm Surge Simulator.
Dr. Chen's work on 3D Hurricane Storm Surge animation has been used by NOAA's National Weather Service (NWS) Tampa Bay Area Office. Demo videos of the project were showcased by the Tampa NWS at the following public outreach events:
For his work on the project, Dr. Chen has received several appreciation e-mails from NOAA's NWS Tampa Bay Area Weather Forecast Office and the Storm Surge Group at NOAA:
On behalf of the NWS Tampa Bay Area Weather Forecast Office - a huge thank you! The video will be an excellent complement to our active outreach program in support of hurricane preparedness focused on storm surge impacts. The video will enable our audiences to see first hand some of the related impacts we have been educating them on for years.
The visual nature of [the video] describing surge resonated with the audience.
Thanks so much for your ongoing support and willingness to work with us toward the development of a new outreach tool.
Dr. Chen and a team of graduate and undergraduate students, developed a new 3D Virtual Reality Visualization system for Integrated Computer Augmented Virtual Environment (I-CAVE) that makes use of Geographic Information System (GIS) data to attest natural disasters such as storm surge. Please visit SCIS web site for details.
The following are relevant papers:
K. Zhang, S.-C. Chen, P. Singh, K. Saleem, and N. Zhao, A 3D Visualization System for Hurricane Storm Surge Flooding, IEEE Computer Graphics and Applications, vol. 26, issue 1, pp. 18-25, Jan.-Feb. 2006.
Keqi Zhang, Jianhua Yan, and Shu-Ching Chen, Automatic Construction of Building Footprints from Airborne LIDAR Data, IEEE Transactions on Geoscience and Remote Sensing, vol. 44, issue 9, pp. 2523-2533, September 2006. (This paper has been cited 262 times by Google Scholar.)
Jianhua Yan, Keqi Zhang, Chengcui Zhang, Shu-Ching Chen and Giri Narasimhan, Automatic Construction of 3D Building Model From Airborne LIDAR Data Through 2D Snake Algorithm, IEEE Transactions on Geoscience and Remote Sensing, Volume 53, Issue 1, pp. 3-14, January 2015.
Maria E. Presa Reyes and Shu-Ching Chen, A 3D Virtual Environment for Storm Surge Flooding Animation, The Third IEEE International Conference on Multimedia Big Data (IEEE BigMM 2017), Laguna Hills, California, USA, pp. 244-245, April 19-21, 2017. (Demo paper)
For a demo of the project, play the videos below.
Thanks to the availability of mobile devices, emergency responders, supporting agencies and even private citizens can capture imagery of disaster events as they unfold. Once the crisis is contained, however, it's a daunting task for emergency managers to collect, organize and integrate disaster event data from multiple sources into incidence command systems where situation reports, incidence action plans, etc. are being held. Therefore, Dr. Chen's group has developed a semi-autonomous system, the Multimedia-Aided Disaster Information Integration System (MADIS), that uses advanced data integration and visual analysis techniques to associate temporal, spatial and other textual features of a disaster event situation report with event images and related text annotations.
The system is developed on Apple's mobile operating system (iOS) and runs on iPad tablets, and it is evaluated by domain experts from the local emergency management department. By interacting with the Miami-Dade Emergency Management (MDEM) personnel through evaluation and exercise activities, the system is constantly being updated by improving user interface experience and back-end support techniques. Feedback from our collaborative partners at MDEM and the potential users suggests that our system will be very useful for emergency managers to gain insight of the situation at the actual disaster scene and to respond swiftly. It is also encouraged to further develop the system into an operational pilot and promote the commercialization of the system for benefiting the whole Emergency Management community.
The system, developed as part of the project A Data Mining Framework for Enhancing Emergency Response Situation Reports with Multi-Agency, Multi-Partner, Multimedia Data, received funding from the Visual Analytics for Command, Control, and Interoperability Environments (VACCINE), the Department of Homeland Security's (DHS) Center of Excellence in Visual and Data Analytics, which was established in July of 2009.
For a demo of the project, play the video below.
Dr. Chen's Business Continuity Information Network (BCIN) project is the first web-based Public-Private Partnership tool that helps the County Business Recovery Program communicate, share information, and collaborate on disaster events with the private sector.
By tracking the status of critical services over multiple counties, the BCIN helps businesses to better assess the impact of a disaster on the community and find resources to recover faster after a disaster. Additionally, it provides a gathering place for businesses to report on the available resources (products/services) they can provide, and helps county partners collect damage report data from county and company sources in order to understand the amount of damage in the business community immediately after a disaster event.
The role of the BCIN in the success of businesses is invaluable. Studies show that about 40% of the companies that closed for three or more days as a result of a hurricane failed within 36 months. If the BCIN helped 5% of the companies in South Florida to speed up their hurricane recovery by one week, it would prevent $220 million of non-property economic losses that would result from that week's closure.
Many businesses have already joined the Network. Some of the participating businesses are listed below:
Since its release, the BCIN has been actively used in many community activities, some of which are listed below:
Dr. Chen has received multiple recognitions as his role has been critical for the success of the BCIN.
The development and implementation of the Business Continuity Information Network (BCIN) has provided the [Miami-Dade County's Business Recovery] program with a key tool necessary to carry out many of its critical functions including real-time data collection and gathering. BCIN is quickly becoming a nationally recognized information sharing tool for use by public-private collaborative programs ... FIU's contribution to the success of Miami-Dade County's Business Recovery Program is truly immeasurable. (Read recognition letter).
Part of a recognized Public-Private Partnership model by the FEMA Private Sector Office is available here.
New U.S.-Japan collaborations bring Big Data approaches to disaster response - NSF Press release 15-029 is available here. United States Senator Bill Nelson's recognition letter to project PI, Dr. Tao Li, is available here. Dr. Chen is the Co-PI of this project.
Click here to visit the Business Continuity Information Network's public service.
Dr. Chen's work on the Coordinated Damage Assessment Application (CDAA) is a reinterpretation of Miami-Dade County's SnapShot Damage Assessment Application.
The CDAA has been made available to the residents of Miami-Dade County, and has assisted in providing accurate situational awareness regarding the impact of a disaster as well as has helped emergency responders plan an appropriate response and recovery.
The CDAA is composed of two components: a public page and a private administration web application, both of which were powered by an application called Snapshot, which was officially used by the Miami-Dade Department of Emergency Management (DEM).
The CDAA is implemented with shared codebase to eliminate as much platform-specific code as possible, modularized code for organized and easily visualized application structure, and lightweight frameworks to abstract several developmental processes. It provides support for assessing either online or offline via PCs, laptops, tablets, and smartphones, and provides an aggregated view of the impact of the disaster.
The goals of this application are to leverage on the community's altruism, provide the community a role and leave them situationally aware, streamline impact assessments, rapidly gather disaster impact information, and efficiently plan for a response and recovery.
From the community's point of view, the CDAA's main roles are to simplify the assessment process for the affected properties and to provide feasible channels for having assessments. On the other hand, the CDAA is implemented for government officials to quickly update the assessment process and to support the ability to easily view and utilize aggregated impact results.
In the paper Affinity Hybrid Tree: An Indexing Technique for Content-Based Image Retrieval in Multimedia Databases published in Proceedings of the IEEE International Symposium on Multimedia (ISM2006), pp. 47-54, December 11-13, 2006, San Diego, CA, USA, Dr. Chen and his PhD student Kasturi Chatterjee proposed the Affinity Hybrid Tree (AH-Tree), a novel indexing and access method to organize large image data sets efficiently and to support popular image access mechanisms like Content-Based Image Retrieval (CBIR) by embedding the high-level semantic image-relationship in the access mechanism as it is.
The AH-Tree combines SpaceBased and Distance-Based indexing techniques to form a hybrid structure that is efficient in terms of computational overhead and fairly accurate in producing query results close to human perception.
The proposed index structure solves the existing problems of introducing high-level image relationships in a retrieval mechanism without going through the pain of translating the content-similarity measurement into feature-level equivalence and yet maintaining an efficient structure to organize the large sets of images.
The paper won the Best Paper Award.
In the paper A Novel Anomaly Detection Scheme Based on Principal Component Classifier, published in Proceedings of the IEEE Foundations and New Directions of Data Mining Workshop, in conjunction with the Third IEEE International Conference on Data Mining (ICDM'03), pp. 172-179, November 19-22, 2003, Melbourne, Florida, USA., Dr. Chen and his colleagues Dr. Mei-Ling Shyu, Dr. Kanoksri Sarinnapakorn, and Dr. LiWu Chang, proposed a novel scheme that uses robust principal component classifier in instrusion detection problems where the training data may be unsupervised.
Assuming that anomalies can be treated as outliers, they constructed an intrusion predictive model from the major and minor principal components of the normal instances. A measure of the difference of an anomaly from the normal instance is the distance in the principal component space. In the paper, they showed that the distance based on the major components that account for 50% of the total variation and the minor components whose eigenvalues are less than 0.20 work well.
They also showed that experiments with KDD Cup 1999 data demonstrated that the proposed method achieves 98.94% in recall and 97.89% in precision with the false alarm rate 0.92%, and outperforms the nearest neighbor method, density-based local outliers (LOF) approach, and the outlier detection algorithm based on Canberra metric.
The paper has been cited 573 times by Google Scholar.
In the paper Video Semantic Event/Concept Detection Using a Subspace-Based Multimedia Data Mining Framework, published in IEEE Transactions on Multimedia, Special Issue on Multimedia Data Mining, Volume 10, Number 2, pp. 252-259, February 2008, Dr. Chen, together with his colleagues and students, proposed a subspace-based multimedia data mining framework for video analysis, specifically video event/concept detection, by addressing two basic issues: semantic gap and rare event/concept detection.
The proposed framework achieves full automation via multimodal content analysis and intelligent integration of distance-based and rule-based data mining techniques. The content analysis process facilitates the comprenhensive video analysis by extracting low-level and middle-level features from audio/visual channels. The integrated data mining techniques effectively address these two basic issues by alleviating the class imbalance issue along the process, and by reconstructing and refining the feature dimension automatically.
The promising experimental peformance on goal/corner event detection and sports/commercials/building concepts extraction from soccer videos and TRECVID news collections demonstrates the effectiveness of the proposed framework.
Additionally, the framework's unique domain-free characteristic indicates the great potential of extending it to a wide range of different application domains.
TRECVID is one of the most famous international video retrieval competitions, conducted by the National Institute of Standards and Technology (NIST), and has held a total of 15 competitions taking place annually since the year 2003. The Ad-hoc Video Search (AVS) task requests the participants to build a video retrieval system within a one-month period, capable of searching for thirty types of videos in a large-scale database. Each type of the videos is described by a query, such as "Find shots of a projection screen" and "Find shots of a truck standing still while a person is walking beside or in front of it."
In the proposed framework, the most recent advances in video processing and deep learning are utilized to understand the contents of the videos. The results outperformed those of several recognized research groups. The development of this system can potentially promote the search engine's capability to understand general descriptions of the video's content without relying on user-created text metadata.
The proposed work won the third place in TRECVID 2018 (AVS task). The FIU News article detailing this achivement can be found here.