Teaching

Image processing

(Year III, 2-nd semester)

English class: Code: IMGP.105.AC.01.04-B8; Lecturer: Sergiu Nedevschi
Romanian class: Code: IMGP.105.AC.01.02-B8 ; Lecturer: Tiberiu Marita


Objectives
The course is designed to introduce students to theoretical concepts and practical issues associated with image processing. The following topics are covered: image preprocessing, image enhancement, image segmentation and analysis. A special effort will be made to develop students' problem solving skills and engineering intuition in the subject area. Upon completion of the course, the student should be knowledgeable and competent in applying the concepts, and should be capable of reading advanced textbooks and research literature in the image-processing field.


Key words
Digital image processing, enhancement, preprocessing, segmentation and analysis, computer vision.


Lecture Curricula

Computer vision and applications. Vision systems structure and functions. Image acquisition systems. Image formation and sensing. Camera model. Stereo-vision basics. Epipolar geometry. Depth computation. Image data representation.

Binary image processing

Single object: geometric properties from image, geometric properties from run-lengths encoding, Multiple objects: connected components, labeling algorithms, contour tracking, and contour approximation. Mathematical image morphology, distance transform, medial axis, thinning.

Grayscale image processing

Mathematical methods for grayscale image processing: Convolution, Fourier transform. Statistic features of the grayscale images. Histogram processing. Noise in grayscale images. Image enhancement: spatial domain methods, frequency domain methods, geometric transformations. Image blurring and sharpening.

Grayscale image segmentation

Region based segmentation: threshold methods, texture based methods. Region growing, region splitting and merging, Wathershading. Edge based segmentation: first order differential methods, second order differential methods. Edge linking, contour closing, contour approximation by straight lines. Hough transform.

Color images. Color models. Color based segmentation.

Textbooks and references
R.C.Gonzales, R.E.Woods, "Digital Image Processing-Second Edition", Prentice Hall, 2002.
E. Trucco, A. Verri, "Introductory Techniques for 3-D Computer Vision", Prentice Hall, 1998.
R. Haralik, L. Shapiro, "Computer and Robot Vision", Addison Wesley, 1993.
I. Pitas, "Digital Image Processing Algorithms", Prentice Hall, 1993.
Y. Shirai, "Three-dimensional Computer vision", Springer-Verlag, 1987.
S. Nedevschi, "Prelucrarea imaginilor si recunoasterea formelor", Ed. Microinformatica, 1997.

 

Pattern recognition systems

(Year IV, 1-st semester)

Romanian class: Code: PREC.105.AC.01.02-B9 ; Lecturer: Sergiu Nedevschi


Objectives
The course is designed to introduce students to theoretical concepts and practical issues associated with pattern recognition. The following topics are covered: model based pattern recognition using statistic and structural approaches. A special effort will be made to develop students' problem solving skills and engineering intuition in the subject area. Upon completion of the course, the student should be knowledgeable and competent in applying the concepts, and should be capable of reading advanced textbooks and research literature in the pattern recognition field.


Key words
Model based recognition, features, constraints, knowledge representation, intermediate representation, statistic pattern matching, structural pattern matching, exact matching, inexact matching.


Lecture Curricula

Model based pattern recognition

The mathematic model of the three-dimensional recognition from depth images.
The mathematic model of three-dimensional recognition from intensity images.
The mathematic model of two-dimensional recognition from intensity images.
Model based recognition problems.
Computational strategies.

Statistic pattern recognition

Review of probability and statistics.
Decision theory.
Linear and quadratic classifiers.
Density estimation.

Structural pattern recognition

Features selection and extraction.
Constraints.
Model and scene representation.
The exact matching.
Search space.
Exhaustive matching methods: Graph isomorphism methods. Relaxation methods. Transforms classifications.
Search space reduction methods: Tree search. Hypothesis generation and checking.

Intermediate representation.

Inexact matching.

Knowledge based recognition

Textbooks and references
Richard O. Duda, Peter E. Hart , David G . Stork, "Pattern Clasification", John Wiley and Sons, 2001
W.E. Grimson, " Object Recognition by Computer: The Role of Geometric Constraints", MIT Press, 1990.
S. Nedevschi, "Prelucrarea imaginilor si recunoasterea formelor", Ed. Microinformatica, Cluj, 1997.

 

Advanced Pattern Recognition Methods

(Year V, 1-st semester)

Romanian class: Code: APRM.105.AC.25.01-M2 ; Lecturer: Sergiu Nedevschi


Lecture Curricula

Mid-level Vision

Segmentation by Clustering
Segmentation by Fitting a Model
Segmentation and Fitting using Probabilistic Methods
Tracking with Linear Dynamic Models

High-level Vision: Geometric Methods

Model-Based Vision
Smooth Surfaces and their Outlines
Aspect Graphs
Range Data

High-level Vision

Probabilistic and Inferential Methods
Finding Templates using Classifiers
Recognition by Relations between Templates
Geometric Templates from Spatial Relations

 

Textbooks and references
D. Forsyth, J. Ponce, Computer Vision - A Modern Approach, Prentice Hall, 2003
E. Trucco, A. Verri, Introductory Techniques for 3-D Computer Vision, Prentice Hall, 1998
R. Gonzales, R. Woods, Digital Image Processing, Addison-Wesley, 2002