 |
Yao-Yi Chiang, Ph.D.Research Scientist,Information Sciences Institute University of Southern California Lecturer, Spatial Sciences Institute University of Southern California Email: yaoyichi_at_gmail_dot_com |
[Curriculum Vitae] [Research] [Projects] [Teaching] [Publications] [Videos] [Awards] [Patents] [Professional Information] [Data Download] [中文網頁] |
|
My general area of research is information integration, with a focus on discovery, extraction, and fusion of spatiotemporal data from heterogeneous sources.
In particular, I am currently interested in geographic information integration (integrating spatial information in geographical space). For example, the techniques that integrate geographic data that are dynamically produced in urban areas from stationary sensors, mobile devices, people, and vehicles.
Keyword cloud from my publications:
If you want to meet me, I am a regular attendee of the following conferences/workshops: the ACM SIGSPATIAL international conference on advances in geographic information systems (ACM GIS), international conference on document analysis and recognition (ICDAR), IAPR international workshop on graphics recognition (GREC), and international conference on pattern recognition (ICPR).
My Ph.D. dissertation (completed in 2010) and prior industry experiences focused on general techniques that enable geographic information systems (GISs) to access geographic information (spatial information in geographical space) “locked” in heterogeneous map images and integrate the geographic information from map images with other spatial data. These techniques exploit location-based information and common map properties to automatically discover map images on the Internet and convert map images into GIS accessible data formats, such as machine editable text and vectors (points, lines, and polygons). The results are spatially registered data (e.g., named road data) that are fused with existing datasets and can be used in spatial analyses. To operationalize the findings of this research, I built a map processing system called Strabo, which has been demonstrated to various U.S. government agencies and research institutions.
Historical Geographic Data Extraction
Recent research efforts have examined techniques for processing historical data to enable chronicled spatiotemporal analysis of demographic changes. In collaboration with Professor Stefan Leyk from the Geography Department at the University of Colorado, we presented the first general technique that minimizes user efforts for extracting geographic data (e.g., locations of roads) from historical maps that have limited graphical quality. This technique is an advancement from previous approaches, which generally rely on intensive manual work and require expert knowledge to fine-tune parameters for the applied techniques and thus are not readily usable by non-expert users.
Instructor, Spatial Databases, University of Southern California (GEOG 582, 2011 Summer and Fall)
Co-instructor (with Craig A. Knoblock), Geospatial Data Integration, University of Southern California (CSCI 599, 2011 Spring) Course Syllabus
Teaching Assistant, Information Integration on the Web, University of Southern California (CSCI 548, 2009 Spring) Course Syllabus
There videos demonstrate our map processing system, Strabo:
Strabo Demonstration: extracting named road vector data from an open source map
[mp4 video download] [avi video download]
The 2nd Best Paper award of the 4th Annual Intelligent Systems Division Graduate Student Symposium (2009)
The 2nd Best Paper award of the 3th Annual Intelligent Systems Division Graduate Student Symposium (2008)
The Viterbi School Doctoral Fellowship (2007 to 2011)
Research scientist@Fetch Technologies
Research scientist@Geosemble Technologies
Software engineer@TLJ Intertech Inc.
Free Hit Counter