UNIVERSITY OF KERALA - Marian Engineering College

B.Tech Comp. Sc. & Engg., University of Kerala 65
08.704 (4) PATTERN RECOGNITION AND SCENE ANALYSIS (ELECTIVE I) 3 – 1 – 0
Module I (16 hours)
Introduction. Probability – Probabilities of events, Random Variables, Joint Distributions and Densities,
Moments of Random Variables, Estimation of Parameters from Samples, Minimum Risk Estimators.
Statistical Decision Making – Bayes' Theorem, Multiple Features, Conditionally Independent Features,
Decision Boundaries, Unequal Costs of Error, Estimation of Error Rates, The Leaving-One-Out Technique,
Characteristic Curves, Estimating the Composition of Populations.
Module II ( 18 hours)
Nonparametric Decision Making – Histograms, Kernel and Window Estimators, Nearest Neighbor
Classification Techniques, Adaptive Decision Boundaries, Adaptive Discriminant Functions, Minimum
Squared Error Discriminant Functions, Choosing a Decision Making Technique. Clustering – Hierarchical
clustering, Partitional Clustering. Artificial Neural Networks – Nets without Hidden Layers, Nets with Hidden
Layers, Back-Propagation Algorithm, Hopfield Nets.
Module III (18 hours)
Processing of Waveforms and Images – Gray Level Scaling Transformations, Histogram Equalization,
Geometric Image Scaling and Interpolation, Smoothing Transformations, Edge Detection, Laplacian and
Sharpening Operators, Line Detection and Template Matching, Logarithmic Gray Level Scaling. Image
Analysis – Scene Segmentation and Labelling, Counting Objects, Perimeter Measurement, Following and
Representing Boundaries, Least Squares and Eigenvector Line Fitting, Shapes of Regions, Morphological
Operations, Texture.
Text Books:
1. Pattern Recognition and Image analysis – E. Gose, R. Johnsonbaugh, S. Jost, , PHI
Reference Books:
1. Pattern Classification – R. O. Duda, P.E. Hart, D.G. Stork, , Wiley India Edition
2. Pattern Classification and Scene Analysis – R.O. Duda and P. E. Hart, John Wiley & Sons, New York, 1973.
3. Syntactic Pattern Recognition and Applications – K.S. Fu, , Prentice Hall, Eaglewood cliffs, N.J.,1982
4. Pattern Recognition, 3 rd Edn – S. Theodoridis, K. Koutroumbas, Elsevier
5. Neural Network for Pattern Recognition – C. M. Bishop, Oxford University Press, New York, 1998
Internal Continuous Assessment (Maximum Marks-50)
25 Marks - Tests (minimum 2)
15 Marks - Assignments (minimum 3) such as home work, problem solving, literature survey, seminar,
term-project, software exercises, etc.
10 Marks - Regularity in the class
University Examination Pattern
PART A: Short answer questions 10 x 4 marks=40 marks
All questions are compulsory. There should be at least three questions
from each module and not more than four questions from any module.
PART B: Descriptive/Analytical/Problem solving questions 3 x 20 marks=60 marks
Candidates have to answer one question out of two or two questions
out of four from each module.
Maximum Total Marks: 100