Code No: 07A71105

SET-1

R07

.in

B.Tech IV Year I Semester Examinations, December-2011 IMAGE PROCESSING AND PATTERN RECOGNITION (BIO-MEDICAL ENGINEERING) Time: 3 hours Max. Marks: 80 Answer any five questions All questions carry equal marks --1.a) b)

Discuss about major problem areas in pattern recognition system design. Explain about Finger print recognition with block diagram.

2.a) b)

Define a decision function. Discuss about linear decision functions in detail.

3.a) b)

What are the advantages of pattern classification using statistical approach? Discuss about Bayes classification algorithm for normal patterns with suitable example. [4+12]

4.a) b)

Define gradient vector and its properties. Discuss about least mean square error algorithm with suitable examples.

5.a)

Distinguish between statistical approach and gradient approach with respect to trainable patterns classifiers. Discuss about the Robbins – Monro algorithm using an illustration. [4+12]

[4+12]

or ld

uW

b)

[8+8]

[4+12]

6.a) b)

What are the steps involved in digital image processing? Explain how a digital image is formed and represented.

7.a)

Discuss about contrast stretching, bit-plane slicing and gray level enhancement techniques. Explain the operation of low pass filter in frequency domain for enhancement of digital image. [8+8]

b)

What are the different derivative operators useful in image segmentation? Explain their role. Give the masks of Roberts, Prewitt, and Sobel edge detectors and compare them. [8+8]

nt

8.a)

Aj

b)

[8+8]

******

Code No: 07A71105

SET-2

R07

1.a) b)

Explain the design concepts and methodologies for automatic pattern recognition. With the help of example explain the character recognition process.

[8+8]

4.a) b)

What are the advantages of gradient operators in the patterns classification? Derive the LMSE algorithm in the pattern classification. [4+12]

5.a) b)

Discuss about Increment-correction algorithm. Compare Increment-correction algorithm with Robbins-Monro algorithm. [12+4]

6.a) b)

Write the applications of DCT[discrete cosign transform]. Develop a 2D - DCT [discrete cosign transform] kernel for N=4.

[4+12]

7.a) b)

Define histogram for a digital image. Explain the histogram equalization procedure with suitable example.

[4+12]

8.a) b)

Explain various methods of detecting discontinuities in an image. What is the role of masks in edge detection and discuss about any two masks for edge detection? [8+8]

Aj

nt

3.a)

or ld

b)

Distinguish between linear decision functions and distance functions. With suitable illustrations discuss about minimum-distance pattern classification. [4+12] Distinguish between deterministic pattern classification and statistical pattern classification. Discuss about Basic perceptron model. [4+12]

uW

2.a) b)

.in

B.Tech IV Year I Semester Examinations, December-2011 IMAGE PROCESSING AND PATTERN RECOGNITION (BIO-MEDICAL ENGINEERING) Time: 3 hours Max. Marks: 80 Answer any five questions All questions carry equal marks ---

******

Code No: 07A71105

SET-3

R07

1.a) b)

What are the fundamental problems involved in pattern recognition system design? Discuss about biomedical applications of pattern recognition. [8+8]

4.a) b)

Discuss the perceptron algorithm using gradient approach. Discuss about least – mean –square –error algorithm.

5. a) b)

Define syntactic pattern recognition. Discuss about the grammar used for syntactic pattern recognition with various definitions involved in the grammar. [4+12]

6.a) b)

Explain the elements of digital image processing system. Discuss about the effects of low spatial – resolution and low gray-level resolution in digital image. [8+8]

7.a) b)

Distinguish between spatial and frequency domain enhancement techniques . Explain the operation of image sharpening filter in frequency domain with proper example. [4+12]

8.a)

What is meant by image compression? And what is the need for image compression? Discuss about variable length coding algorithm with a suitable example.

or ld

3.a) b)

What is meant by cluster seeking? Discuss about a cluster seeking algorithm based on Euclidean distance concept. [4+12] List out the advantages of deterministic pattern classification methods. Discuss about Reward –punishment concept with suitable examples. [4+12]

Aj

b)

[8+8]

uW

nt

2.a) b)

.in

B.Tech IV Year I Semester Examinations, December-2011 IMAGE PROCESSING AND PATTERN RECOGNITION (BIO-MEDICAL ENGINEERING) Time: 3 hours Max. Marks: 80 Answer any five questions All questions carry equal marks ---

******

[8+8]

Code No: 07A71105

SET-4

R07

.in

B.Tech IV Year I Semester Examinations, December-2011 IMAGE PROCESSING AND PATTERN RECOGNITION (BIO-MEDICAL ENGINEERING) Time: 3 hours Max. Marks: 80 Answer any five questions All questions carry equal marks --1.a) b)

Explain the principle of pattern recognition With the help of neat block diagram describe the simple pattern recognition model. [8+8]

2.a) b)

Define performance index. Discuss the procedure of k-means algorithm with a simple illustration.

3.a) b)

What is meant by training in pattern classification? Discuss the basic perceptron approach with corrections.

4.a)

Distinguish between gradient approach and other approaches in patterns. classification. Discuss about multi category classification algorithm with suitable example. [4+12]

6.a)

or ld

5.a) b)

Discuss about recognition grammar for syntax-based recognition . Explain the syntax – Directed recognition algorithm based on graphs.

uW

b)

[4+12]

[4+12]

[4+12]

What are the advantages of wavelet transforms when compared to frequency transforms? b) Develop a successive doubling FFT algorithm for N=8. [4+12] What is the significance of histogram of digital image in enhancement? Explain about histogram specification algorithm in a digital image.

8.a)

What are the various redundancies available in digital images? Explain about them. Discuss about Huffman coding algorithm with a suitable example. [6+10]

nt

7.a) b)

Aj

b)

******

[4+12]