Code No:

05411101

Set No. 1

JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD

IV B.Tech I-Sem I Mid-Term Examinations, Aug/Sept. 2009
IMAGE PROCESSING AND PATTERN RECOGNITION
Objective Exam
Name: ______________________________ Hall Ticket No.
A

Answer All Questions. All Questions Carry Equal Marks.Time: 20 Min. Marks: 20.
I

Choose the correct alternative:

1.

Character recognition utilizes the following concept:

a. Heuristic b. Linguistic c. Syntactic
d. Membership roster

2.

The conditional average risk / function in a Bayes Classifier is given by r:

a. Lij p ( wi / x)
b. Lij p( wi / x) c. Lij p ( x / wi ) d. L ji p ( wi / x )

3.

For a function f the negative of a gradient points in the direction of:

a. Max rate of decrease of f
b. Min rate of decrease of f
c. Max rate of increase of f
c. Min rate of increase of f

4.

The minimum number of decision functions for M pattern classes using case 2 is:
a. M b. M+1
c. M(M-1)
d. M(M-1)/2

5.

The LMSE algorithm converges when

a. w > 0
b. X 0
c. Xw > 0

d. Xw > 0

6.

Which of the following does not constitute in the design of automatic pattern recognition model of
a system:
[
]
a. Observation
b. Sensing
c. Feature selection
d. Preprocessing

7.

In a two-pattern class case, a linear decision function, d(x), will classify a sample set x to the
pattern class when:
[
a. d(x)>0 or d(x) < 0 b. d(x)<0 or d(x)>=0 c. d(x)=0
d. d(x) is not used

8.

The classifier that minimizes the total expected loss is called:

a. Optimal classifier b. Likelihood classifier c. Bayes Classifier

9.

In LMSE algorithm, X + + is called

a. Least mean b. Squared error

10.

[
d. Best classifier
[

c. Generalized Inverse

d. Gradient

In the minimum distance algorithm, the decision function d(x) assigns x to wi if

[
]
a. d i ( x ) = d j ( x )
b. d i ( x ) < d j ( x )
c. d i ( x ) > d j ( x )
d. d i (x ) cannot help

Cont2

Code No:

05411101

:2:

Set No. 1

II

Fill in the blanks:

11.

The recognition function rij(x) represents a __________________.

12.

When a priori probabilities are not available ____________ criterion offers an alternative approach.

13.

Minimum distance classification is also known as _________________ matching.

14.

Least mean square error algorithm is also known as the ____________________ algorithm.

15.

A ___________ is a category determined by some given common attributes

16.

Test of separability is an important feature in ____________________ algorithm.

17.

The reward punishment concept is also known as ________________ algorithm.

18.

Pattern classification by distance functions will yield satisfactory results only when patterns
have ________________ properties

19.

|e| represents the _______________ of the vector e

20.

In the gradient descent algorithm, w is incremented in the _________________ gradient of

J(w,x)

-oOo-

Code No:

05411101

Set No. 2

JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD

IV B.Tech I-Sem I Mid-Term Examinations, Aug/Sept. 2009
IMAGE PROCESSING AND PATTERN RECOGNITION
Objective Exam
Name: ______________________________ Hall Ticket No.
A

Answer All Questions. All Questions Carry Equal Marks.Time: 20 Min. Marks: 20.
I

Choose the correct alternative:

1.

The minimum number of decision functions for M pattern classes using case 2 is:
a. M b. M+1
c. M(M-1)
d. M(M-1)/2

2.

The LMSE algorithm converges when

a. w > 0
b. X 0
c. Xw > 0

d. Xw > 0

3.

Which of the following does not constitute in the design of automatic pattern recognition model of
a system:
[
]
a. Observation
b. Sensing
c. Feature selection
d. Preprocessing

4.

In a two-pattern class case, a linear decision function, d(x), will classify a sample set x to the
pattern class when:
[
a. d(x)>0 or d(x) < 0 b. d(x)<0 or d(x)>=0 c. d(x)=0
d. d(x) is not used

5.

The classifier that minimizes the total expected loss is called:

a. Optimal classifier b. Likelihood classifier c. Bayes Classifier

6.

In LMSE algorithm, X + + is called

a. Least mean b. Squared error

[
d. Best classifier
[

c. Generalized Inverse

d. Gradient

7.

In the minimum distance algorithm, the decision function d(x) assigns x to wi if

[
]
a. d i ( x ) = d j ( x )
b. d i ( x ) < d j ( x )
c. d i ( x ) > d j ( x )
d. d i (x ) cannot help

8.

Character recognition utilizes the following concept:

a. Heuristic b. Linguistic c. Syntactic
d. Membership roster

9.

The conditional average risk / function in a Bayes Classifier is given by r:

a. Lij p ( wi / x)
b. Lij p( wi / x) c. Lij p ( x / wi ) d. L ji p ( wi / x )

10.

For a function f the negative of a gradient points in the direction of:

a. Max rate of decrease of f
b. Min rate of decrease of f
c. Max rate of increase of f
c. Min rate of increase of f

Cont2

Code No:

05411101

:2:

Set No. 2

II

Fill in the blanks:

11.

Least mean square error algorithm is also known as the ____________________ algorithm.

12.

A ___________ is a category determined by some given common attributes

13.

Test of separability is an important feature in ____________________ algorithm.

14.

The reward punishment concept is also known as ________________ algorithm.

15.

Pattern classification by distance functions will yield satisfactory results only when patterns
have ________________ properties

16.

|e| represents the _______________ of the vector e

17.

In the gradient descent algorithm, w is incremented in the _________________ gradient of

J(w,x)

18.

The recognition function rij(x) represents a __________________.

19.

When a priori probabilities are not available ____________ criterion offers an alternative approach.

20.

Minimum distance classification is also known as _________________ matching.

-oOo-

Code No:

05411101

Set No. 3

JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD

IV B.Tech I-Sem I Mid-Term Examinations, Aug/Sept. 2009
IMAGE PROCESSING AND PATTERN RECOGNITION
Objective Exam
Name: ______________________________ Hall Ticket No.
A

Answer All Questions. All Questions Carry Equal Marks.Time: 20 Min. Marks: 20.
I

Choose the correct alternative:

1.

Which of the following does not constitute in the design of automatic pattern recognition model of
a system:
[
]
a. Observation
b. Sensing
c. Feature selection
d. Preprocessing

2.

In a two-pattern class case, a linear decision function, d(x), will classify a sample set x to the
pattern class when:
[
a. d(x)>0 or d(x) < 0 b. d(x)<0 or d(x)>=0 c. d(x)=0
d. d(x) is not used

3.

The classifier that minimizes the total expected loss is called:

a. Optimal classifier b. Likelihood classifier c. Bayes Classifier

4.

In LMSE algorithm, X + + is called

a. Least mean b. Squared error

[
d. Best classifier

5.

In the minimum distance algorithm, the decision function d(x) assigns x to wi if

[
a. d i ( x ) = d j ( x )
b. d i ( x ) < d j ( x )
c. d i ( x ) > d j ( x )
d. d i (x ) cannot help

6.

Character recognition utilizes the following concept:

a. Heuristic b. Linguistic c. Syntactic
d. Membership roster

7.

The conditional average risk / function in a Bayes Classifier is given by r:

a. Lij p ( wi / x)
b. Lij p( wi / x) c. Lij p ( x / wi ) d. L ji p ( wi / x )

8.

For a function f the negative of a gradient points in the direction of:

a. Max rate of decrease of f
b. Min rate of decrease of f
c. Max rate of increase of f
c. Min rate of increase of f

9.

The minimum number of decision functions for M pattern classes using case 2 is:
a. M b. M+1
c. M(M-1)
d. M(M-1)/2

10.

The LMSE algorithm converges when

a. w > 0
b. X 0
c. Xw > 0

c. Generalized Inverse

d. Gradient

d. Xw > 0

Cont2

Code No:

05411101

:2:

Set No. 3

II

Fill in the blanks:

11.

Test of separability is an important feature in ____________________ algorithm.

12.

The reward punishment concept is also known as ________________ algorithm.

13.

Pattern classification by distance functions will yield satisfactory results only when patterns
have ________________ properties

14.

|e| represents the _______________ of the vector e

15.

In the gradient descent algorithm, w is incremented in the _________________ gradient of

J(w,x)

16.

The recognition function rij(x) represents a __________________.

17.

When a priori probabilities are not available ____________ criterion offers an alternative approach.

18.

Minimum distance classification is also known as _________________ matching.

19.

Least mean square error algorithm is also known as the ____________________ algorithm.

20.

A ___________ is a category determined by some given common attributes

-oOo-

Code No:

05411101

Set No. 4

JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD

IV B.Tech I-Sem I Mid-Term Examinations, Aug/Sept. 2009
IMAGE PROCESSING AND PATTERN RECOGNITION
Objective Exam
Name: ______________________________ Hall Ticket No.
A

Answer All Questions. All Questions Carry Equal Marks.Time: 20 Min. Marks: 20.
I

Choose the correct alternative:

1.

The classifier that minimizes the total expected loss is called:

a. Optimal classifier b. Likelihood classifier c. Bayes Classifier

[
d. Best classifier

2.

In LMSE algorithm, X + + is called

a. Least mean b. Squared error

d. Gradient

[
c. Generalized Inverse

3.

In the minimum distance algorithm, the decision function d(x) assigns x to wi if

[
]
a. d i ( x ) = d j ( x )
b. d i ( x ) < d j ( x )
c. d i ( x ) > d j ( x )
d. d i (x ) cannot help

4.

Character recognition utilizes the following concept:

a. Heuristic b. Linguistic c. Syntactic
d. Membership roster

5.

The conditional average risk / function in a Bayes Classifier is given by r:

a. Lij p ( wi / x)
b. Lij p( wi / x) c. Lij p ( x / wi ) d. L ji p ( wi / x )

6.

For a function f the negative of a gradient points in the direction of:

a. Max rate of decrease of f
b. Min rate of decrease of f
c. Max rate of increase of f
c. Min rate of increase of f

7.

The minimum number of decision functions for M pattern classes using case 2 is:
a. M b. M+1
c. M(M-1)
d. M(M-1)/2

8.

The LMSE algorithm converges when

a. w > 0
b. X 0
c. Xw > 0

d. Xw > 0

9.

Which of the following does not constitute in the design of automatic pattern recognition model of
a system:
[
]
a. Observation
b. Sensing
c. Feature selection
d. Preprocessing

10.

In a two-pattern class case, a linear decision function, d(x), will classify a sample set x to the
pattern class when:
[
a. d(x)>0 or d(x) < 0 b. d(x)<0 or d(x)>=0 c. d(x)=0
d. d(x) is not used

Cont2

Code No:

05411101

:2:

Set No. 4

II

Fill in the blanks:

11.

Pattern classification by distance functions will yield satisfactory results only when patterns
have ________________ properties

12.

|e| represents the _______________ of the vector e

13.

In the gradient descent algorithm, w is incremented in the _________________ gradient of

J(w,x)

14.

The recognition function rij(x) represents a __________________.

15.

When a priori probabilities are not available ____________ criterion offers an alternative approach.

16.

Minimum distance classification is also known as _________________ matching.

17.

Least mean square error algorithm is also known as the ____________________ algorithm.

18.

A ___________ is a category determined by some given common attributes

19.

Test of separability is an important feature in ____________________ algorithm.

20.

The reward punishment concept is also known as ________________ algorithm.

-oOo-