Sibel Tari, Professor (Email: stari[at]metu[dot]edu[dot]tr) Computer Engineering Department Middle East Technical Univ.			ANNOUNCEMENTS I am on sabbatical at TU Munich, Scientific Computing with a funding from Alexander von Humboldt Foundation		TEACHING (2003-Present) CENG 382 Dynamic Systems with Feedback CENG 483 Introduction to Computer Vision CENG 583 Computational Vision CENG 566 Image Processing (1998-2003) CENG 373 Operational Mathematics ES 303 Statistical Methods (at Dept. Eng. Science) ES 305 Numerical Methods ES 509 PDE's in Comp. Vision and Image Proc.

	RESEARCH	PROJECTS	STUDENTS	DEVIN My Son
	DEGREES
	BS in Computer Science and Engineering, Hacettepe University, 1989 PhD in Computer Vision (Interdisciplinary Degree), Northeastern University, 1997 Advisor: Jayant Shah Docent Degree in Theoretical Foundations of Computer Science, 2000
	SELECTED PUBLICATIONS/RESEARCH
	My research interest is in cognitive and computational aspects of shape perception. Representative Publications: Shape Skeletons Disconnected Skeleton: Shape at its Absolute Scale. IEEE Transactions on PAMI, Vol.30, 2188-2203. pdf (with C. Aslan, A. Erdem, E. Erdem ) An Axis-Based Representation for Recognition. ICCV 2005, 1339-13346. pdf (with C. Aslan ) Local Symmetries of Shapes in Arbitrary Dimension. ICCV 1998, 1123-1128. (with Jayant Shah) Nested Local Symmetry Set. Computer Vision Image Understanding, Vol.79, 267-281. pdf (with Jayant Shah) Extraction of Shape Skeletons from Grayscale Images. Computer Vision Image Understanding, Vol.66, 133-147. pdf (with Jayant Shah and H. Pien) Computationally Efficient Shape Analysis via Level Sets. Proceedings of the. IEEE-SIAM Workshop Math. Methods in Biomedical Image Analysis 1996, 234-244. (with Jayant Shah and H. Pien) Parts, Distance Functions Extracting Parts of 2D Shapes Using Local and Global Interactions Simultaneously. In Handbook of Pattern Recognition and Computer Vision 2010. pdf Hierarchical Shape Decomposition via Level Sets. International Symposium Mathematical Morphology 2009. pdf Components of the Shape Revisited. AAAI Fall Symposium. pdf (with B. Burgeth, I. Tari) Mumford/Shah, Non-Linear Diffusion PDEs From a Non-local Ambrosio-Tortorelli to a Randomized Hierarchy Tree. In: Scale Space and Variational Methods in Computer Vision, LNCS 6667 (Springer), 267Ñ271. (with M. Genctav) pdf Mumford-Shah Regularizer with Contextual Feedback. Journal of Mathematical Imaging and Vision Vol.33, 67-84. (with E. Erdem ) pdf Mumford-Shah Regularizer with Spatial Coherence. In: Scale Space and Variational Methods in Computer Vision, Series: Image Processing, Computer Vision, Pattern Recognition, and Graphics, 4485. Eds: Sgallari, F., Murli, A. and Paragios, N. (Springer Berlin/Heidelberg), 545-555. (with E. Erdem, A. Sancar ) Minimally Stochastic Schemes for Singular Diffusion Equations. In: Image Processing Based on Partial Differential Equations, Eds: X-C Tai, K-A Lie, TF Chan and S Osher. (Springer Berlin/Heidelberg), 325-339. (with B. Burgeth, J. Weickert) Extraction of Shape Skeletons from Grayscale Images. Computer Vision Image Understanding, Vol.66, 133-147. pdf (with Jayant Shah and H. Pien) Computationally Efficient Shape Analysis via Level Sets. Proceedings of the. IEEE-SIAM Workshop Math. Methods in Biomedical Image Analysis 1996, 234-244. (with Jayant Shah and H. Pien) Applied Work in Computational Design Weighted Shapes for Embedding Perceived Holes Environment and Planning, accepted (with H. Keles and M. Ozkar) -- due to Publisher's policy pdf available via email (the Publisher does not support Open Choice). Embedding Shapes without Predefined Parts, Environment and Planning, Vol. 37, 664-681. pdf (with M. Ozkar and H. Keles) Distance Transforms via Elliptic PDEs to be added Applications of Skeletons, Shape Matching etc to be added
	SELECTED PROJECTS
	As Principal Investigator
	(2005-2008) Incorporation of Prior and Contextual Information into Computer Vision - TUBITAK Support link (temporarily unavailable) (2008-2010) Shape Grammars for Analysis of Visual Thinking and Design - TUBITAK Support link (temporarily unavailable)
	As Researcher
	Anatomic Morphometric Segmentation of Brain Images (NIH Support PI: Prof. D. Kennedy of Harvard Medical School) Part of Human Brain Project
	STUDENTS
	PhD Degree
	Aykut Erdem Current Position: Assistant Prof., Hacettepe University PhD Thesis: Category Knowledge, Skeleton-Based Shape Matching and Shape Classification, Sep 2004-Oct 2008 Erkut Erdem Current Position: Current Position: Inst. Dr., Hacettepe University PhD Thesis: Simultaneous Bottom up/Top down Processing in Early and Mid Level Vision, Sep 2004-Nov 2008 Hacer Yalim Keles Current Position: Research Director, Kolektif Bilisim PhD Thesis: Part Embedding for Shape Grammars, Sep 2007-July 2010
	MS Degree
	Cagri Aslan Current Position: Microsoft Corp. Seattle, WA MS Thesis: Disconnected Skeleton, May 2005 pdf thesis Emre Baseski MS Thesis: Context Sensitive Matching of Skeletal Trees, June 2006 > Aysun Sancar MS Thesis: Edge Preserving Smoothing with Directional Consistency, May 2007 Sezen Erdem MS Thesis: Motion Analysis via Axis Based Representation, Sep 2007 Murat Genctav MS Thesis: Nov 2010
	MIND: Mathematical Imaging Non-linear Diffusion
	MIND Group is conducting research in Mathematical Imaging and Nonlinear Diffusion. Mind Page is under construction.