Jawaharlal Nehru Technological University Kakinada 2010-2nd Sem M.Tech Computer Science & Engineering Supplementary s Pattern Recognition and Image Processing (CSE-CS Engineering) - Ques

Subject Code: B5805
M.Tech II Semester Supplementary exams June 2010
trend Recognition and Image Processing
(CSE-CS Engineering)
Time: three Hours Max Marks: 60
ans any 5 ques. All ques. carry EQUAL marks


1. a) discuss about a practical PR system development giving details about
i) Training ii) Data Sets
b) discuss when Bayesian classifier becomes Euclidean distance classifier. Derive a
decision surface for a 2 group issue in such a case.
2. a) discuss about discriminant functions for normal density for various cases.
b) provide a detailed account on Bayes decision theory.
3. a) What is parameter estimation problem?. Why do we need? discuss about Maximum
likelihood estimation.
b) Write a brief note on parameter estimation.
4. define K-Means algorithm. How do we measure performance of a clustering
algorithm. Divide the subsequent 2-D data into 2 groups using K-Means. Assume,
first two samples as initial cluster centers.
X1 20 20
X2 21 20
X3 25 25
X4 27 35
X5 33 35
X6 29 33
X7 34 23
X8 22 33
5. a) define about the 3 basic issues for hidden Markov model in detail.
b) What is discrete time markov process and discuss about extensions to hidden
Markov models.
6. a) provide and discuss of speech recognition applications using continuous hidden
Markov models.
b) explain in detail about continuous observation densities.
7. a) discuss in detail about fundamental steps in digital image processing.
b) discuss gray level and transformation and histogram processing in detail.
8. a) discuss about crack edge detection in detail.
b) explain about thresholding in hierarchical data structures.

Earning:   Approval pending.