Dorothea Blostein
Professor, School of Computing at Queen's University in Kingston, Ontario.

Dorothea Blostein
School of Computing
Goodwin Hall 720
Queen's University
Kingston, Ontario, Canada
K7L 2N8
(613) 533-6537

Dr. Blostein received computer science degrees from the University of Illinois (B.Sc. in 1978, Ph.D. in 1987) and Carnegie Mellon University (M.Sc. in 1980). Since 1988 she has been a faculty member in the School of Computing at Queen's University.

New Research Area: Fascial Network Computing

I am starting work on a novel form of bioinspired computing -- Fascial Network Computing. Fascia, also called connective tissue, forms a bodywide network that provides structural support, protection, shock absorption and elastic recoil. Fascial tissue exhibits a great diversity of characteristics, ranging from stiff cable-like structures such as tendons to supple sheets such as the superficial fascia under the skin. Fascial tissue is adaptive: it changes characteristics in response to the demands placed on it. For example, chronic sitting in a hunched position causes the fascia in the back of neck to become stiffer: the resulting increase in structural support reduces the muscular effort needed to maintain a forward head position, but comes at the expense of reduced neck mobility.
     I propose simulating a Fascial Network as an adaptive tensegrity model. Tensegrity (tensional integrity) is a structural principle popularized by Buckminster Fuller in which isolated components under compression are held in place by a network of components under tension. Tensegrity structures are both strong and flexible due to the dynamic interplay between tension and compression forces. In Fascial Network Computing, the tensioned members provide an abstract representation of tension in fascia (see Fascia: The Tensional Network of The Human Body), with the compressed members representing bone as well as hydrostatic pressure (see The diversity of hydrostatic skeletons). During training of a Fascial Network, simulated external load patterns are repeatedly applied to the network. Adaptation rules dictate how the stiffness of a particular fascial element changes in response to the local load it experiences during training; this mimics the adaptive properties displayed by fascia in living organisms.
     Revealing analogies can be drawn with Neural Network Computing, a well-established form of bioinspired computing that has been successfully used in many pattern recognition applications. The pattern of stiffnesses in a trained Fascial Network constitutes a type of distributed memory, analogous to the distributed memory formed by the pattern of connection strengths in a trained Neural Network. Both Neural and Fascial Networks exhibit emergent properties such as global response to local injury. Fascial Network Computing offers insight into the structural responsiveness of a biological system, an intriguing complement to the neural responsiveness modeled by Neural Network Computing.
     My students and I are beginning investigation of Fascial Network Computing as a topic within computer science: a novel form of bioinspired computation. The longer-term goal is to increase understanding of the biological mechanisms that maintain homeostasis in a fascial network. This requires collaboration with physiologists, clinicians, manual therapists, mathematicians and mechanical engineers. Increased understanding of network homeostasis is essential for improving treatment of common conditions such as lower back pain, whiplash and concussion. Research on the fascial aftereffects of concussion complements extensive existing studies on the neural aftereffects of concussion. If successful, this work has potential application in post-concussion rehabilitation and in estimating future concussion proneness of a patient.

Here are a few relevant links. (See Tom Flemon's Resources page for a more comprehensive list.)

Research in Graphics Recognition and Document Classification

Dr. Blostein has a long-standing research program in pattern recognition, document analysis, and document classification. The main research goal is to smooth the interface between paper and electronic versions of documents. Dr. Blostein and her students have worked on developing computer technology to read, write, and edit diagram notations such as music notation, math notation, maps, schematics, and architectural drawings. The text portions of scanned documents can be analyzed with OCR (Optical Character Recognition), but further processing is required to extract the information contained in diagram notations. Dr. Blostein and her students investigate the use of techniques such as graph transformation and tree transformation, in order to construct an interpretation of the 2D arrangement of symbols in a diagram. Other projects include classification of documents (applied to biomedical documents and software engineering documents), and the use of internet searches to validate document recognition results.
Of special interest to Dr. Blostein is exploration and exploitation of the relationship between diagram recognition and diagram generation. Technology for generation is far ahead of technology for recognition: diagram generation software outperforms diagram recognition software; speech generation outperforms speech understanding; computer graphics is ahead of computer vision. There are research opportunities here!

Dr. Blostein has coauthored Lime, an editor for music notation. Both Macintosh and IBM PC versions are available for free trial use.

Dr. Blostein was a plenary speaker at CICM 2009. She was Chair of GREC2001, the Fourth IAPR International Workshop on Graphics Recognition, Kingston, Ontario, in September, 2001.

Selected Publications

Classification and Mining of Software Engineering Documents

Classification of Biomedical Documents

Holographic Reduced Representations

Formulating and Evaluating Recognition Algorithms using the Recognition Strategy Language

Validation and Performance Evaluation of Document Recognition Algorithms

Surveys on Topics in Diagram Recognition and Document Classification

Recognition of Mathematics Notation Using Tree Transformation

Graph Transformation and Application to Diagram Recognition

Lime Music Notation Software

Computer Vision

Visiting Researchers

Oleg Golubitsky, Postdoctoral Fellow, January 2005 to August 2006.
Research topic: structural representations for document recognition. This builds on existing work with Lev Goldfarb on ETS (Evolving Transformation Systems).
Sergei Levashkin, visiting sabbaticant, August 2004 to August 2005.
Research on recognition of cartographic maps.
Gulila Adongbieke, visiting professor from Xinjiang University, P.R. China.
Research May-Oct 2003 on classification of documents, using linguistic methods.
Clemens Oertel, graduate exchange student.
Research Jan-Aug 2003 on external verification of the results produced by business-card recognition.
Li Zhifeng, visiting researcher from Chinese HanWang Company, Beijing, P.R. China.
Research Oct 2001-Jan 2002 on character recognition and document analysis.

Research Supervision

Current Graduate Students

Perry BhandalStarting MSc May 2015Extensions to the NASA Tensegrity Robotics Toolkit

Theses Supervised

Xingcheng CaiM.Sc. ThesisA Prototype Helmet Fitting System for Concussion Protection (co-supervised with Dr. Fraser Saunders) January 2015
Slava JdanovM.Sc. thesis A Tensegrity Based Structure Optimization Framework August 2014
Adrian MuresanM.Sc. thesis Modeling Tensegrity Systems via Potential Energy Minimization August 2014
Steven ThomasPh.D. thesis Mining Unstructured Software Repositories Using IR Models (co-supervised with Ahmed Hassan) December 2012
Matthew Brian KellyM.Sc. thesis Evaluation of Melody Similarity Measures September 2012
Matthew Alexander KellyM.Sc. thesis Advancing the Theory and Utility of Holographic Reduced Representations (co-supervised with Douglas Mewhort) August 2010
Yin LamM.Sc. thesis Comparing Naive Bayes Classifiers with Support Vector Machines for Predicting Protein Subcellular Location Using Text Features (co-supervised with Hagit Shatkay) June 2010
Andrew SeniukM.Sc. thesis Pen-Chant: Acoustic Emissions of Handwriting and Drawing August 2009
Adrien LapointeM.Sc. thesis Issues in Performance Evaluation of Mathematical Notation Recognition Systems May 2008
Steven (Jianhui) ChenM.Sc. thesis A Wavelet-based Approach to the Classification of Remotely Sensed Images: A Comparison of Different Feature Sets in an Urban Environment (co-supervised with DongMei Chen, Geography) January 2007
Nicole MitchellM.Sc. thesis Music Similarity Metrics: Recognizing tempo, Transposition, Ornamentation, and Accentuation properties January 2007
Marcus MillerM.Sc. thesis A Structured Approach to Object Segmentation in Aeronautical Charts May 2006
Ling ZhangM.Sc. thesis Fuzzy Logic Approach to Recognition of Mathematical Notation February 2005
Richard Zanibbi Ph.D. thesis A Language for Specifying and Comparing Table Recognition Strategies (co-supervised with Jim Cordy) December 2004
Nawei ChenM.Sc. thesis Exploring a Space of Document Image Classifiers December 2004
Yang LiM.Sc. thesis Asymmetric Graph Matching for Registering Satellite Images to Road Maps (co-supervised with Purang Abolmaesumi) April 2004
Sean ChenM.Sc. thesis A Multiscale Domain-Independent Algorithm for Document Image Segmentation July 2003
Hanaa BarakatM.Sc. thesis Training with Positive and Negative Data Samples: Effects on a Classifier for Hand-Drawn Geometric Shapes May 2001
Ed Lank Ph.D. thesis Retargetable On-Line Recognition of Diagram Notations March 2001
Medha Shukla SarkarPh.D. thesis GXL - A Graph Transformation Language with Scoping and Graph Parameters (co-supervised with Jim Cordy) August 2000
Jianping WuM.Sc. thesis Bayesian Estimation of Stereo Disparity from Phase-Based Measurements (co-supervised with David Fleet) March 2000
Richard ZanibbiM.Sc. thesis Recognition of Mathematics Notation via Computer Using Baseline StructureJanuary 2000
Ben GatienM.Sc. thesisSegmentation of Hand-Written Documents Using Minimum Spanning TreesJuly 1997
Mark RhodenizerM.Sc. thesisAutomatic Extraction of Features from Line DrawingsJuly 1997
Hoda FahmyPh.D. thesis Reasoning in the Presence of UncertaintyMarch 1995
Ann GrbavecM.Sc. thesis Recognition of Mathematics Notation Using Graph RewritingJanuary 1995
Anton DriesseM.Sc. thesis Tempo Tracking in Real TimeJune 1992
Hoda FahmyM.Sc. thesis A Graph-Grammar Approach to High-level Music Recognition September 1991
Guylaine CantinM.Sc. thesis The Use of Function-Tables to Specify Complex Algorithms (co-supervised with David Parnas) July 1991

Undergraduate Research Assistants

Sylvester Chiang fall 2013 Validation methods for brain/skull models
Stephanie Huynh summer 2013 Software engineering issues in biomedical computing
Xingcheng Cai summer 2013 Vertebra pose estimation for scoliosis assessment
Slava Jdanov summer 2012 Adaptive tensegrity networks
Yan Lam summer 2007 Datasets of biomedical documents for classifier testing and training
Shauna O'Shea summer 2004
summer 2005
Validation of OCR results on business cards
Adam Bodnar summer 2002
summer 2003
Recognizing text on scanned business cards
Tim Collier summer 2003 Recognizing text on scanned business cards
Nawei Chen summer 2003 Classification of scanned documents
David Tausky summer 2002
summer 2001
Recognition of road networks in topographic maps
Recognition of matrices in math notation
Alvin Jugoon summer 2002
summer 2001
Recognition of business cards
User interface for recognition of UML notation
Arlis Rose summer 2002
summer 2001
Recognition of business cards
User interfaces for diagram recognition
Sean Chen summer 2001
summer/fall 2000
Recognition of handwritten UML notation
Randall Tuesday summer 2001 Graph display; animation of graph algorithms
Jeremy Hussel summer 2001 User interface for recognition of UML notation
Jeb Thorley summer/fall 2000 Recognition of handwritten UML notation
Nick Willan summer 1999 Recognition of handwritten math notation
Richard Zanibbi summer 1998 Visual Language work applied to math recognition
David Kidston summer 1995 Graph rewriting with the PROGRES system
Dwi Faulus summer 1993, 1994 Symbol-recognition for music notation
Hoda Fahmy summer 1989, 1990 High-level recognition of music notation


Current Courses
CISC 859 Pattern Recognition. Fall 2014, 2013, 2011, 2008 and previous years
CISC 324 Operating Systems. Winter 2015, 2014, 2013, 2012, 2009 and previous years

Other Courses I have Taught
CISC 365 Algorithms I. Fall 2010. Also Winter 2003 and previous years.
CISC 124 Introduction to Computing Science II. Fall 2002.
    Here is Java image manipulation code used in this course. This code can be used as a starting point by anyone wishing to write image manipulations in Java.

CISC 352 Artificial Intelligence. Fall 2003.
CISC 221 and CISC 231. Computer Architecture. Winter 1997 and previous years.