COLLEGE OF ENGINEERING KIDANGOOR

B.Tech. Degree VI Semester Model Examination

Time: 3Hrs Max.Marks: 100

Ans wers

I. a) NEIGHBORS

E S . I N
KTU NOT
ADJACENCY

Downloaded from Ktunotes.in

I b) Image formation model

E S . I N
KTU NOT

Downloaded from Ktunotes.in

I c)

E S . I N
KTU NOT

Downloaded from Ktunotes.in

II.a)

E S . I N
NOT
Toeplitz matrix

KTU

Downloaded from Ktunotes.in

 E S . I N
KTU NOT
Circulant matrix

Downloaded from Ktunotes.in

II b)

Spatial shift property

E S . I N
KTU NOT

Periodicity property

Downloaded from Ktunotes.in

 E S . I N
II c)

KTU NOT

III a)

Downloaded from Ktunotes.in

 E S . I N
NOT
III b)

KTU

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT
IVa)

Log transformations, image negatives, histogram equalization , specification, image addition subtraction,
averaging, sharpening.

Histogram

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

IV b)

E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT
Homomorphic filtering approach for image enhancement

V a)

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

V b)

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT
VI a)

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

VI b)

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
NOT
VII a)

KTU
Region based segmentation

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

VII b)

In this , points are linked by determining first if they lie on a curve of specified shape.

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

VII c)

Global Thresholding

E S . I N
KTU NOT
Adaptive Thresholding

Histogram based threshold selection

Downloaded from Ktunotes.in

VIII a)

The text to be encoded consists of 11 characters. As a result, we get the following relative occurrences:

First, we combine the rarest two symbols into a new one and sum up the two relative occurrences
(1/11+2/11=3/11). They will form the deepest branch in the tree. We get a new list of symbols and
again, we combine the two rarest symbols into another node of the tree. Note that in some cases we
might get two not yet connected branches of the tree. However, not in this case:

E S . I N
KTU NOT
Now, we only need to assign either 1 or 0 to each branch and get the optimal Huffman code:

Note that the Huffman tree always produces a prefix-tree because the symbols are only attached to the
leaves. Thus, no symbol can be a (leading) bit pattern of another one. The message MISSISSIPPI would
be coded as: 111 10 0 0 10 0 0 10 110 110 10

VIII b)

JPEG is accepted as an international image compression standard.JPEG (stands for Joint Photographic
Experts Group).

Three different coding systems: baseline coding system; extended coding system, and lossless
independent coding system

1) Baseline coding system involves

• DCT-based: 8x8 block-transform
• Thresholding-based quantization
• Zig-zag based reordering to obtain 1-D sequence of quantized coefficients
Downloaded from Ktunotes.in
 • Huffman coding of AC coefficients
• Difference coding of DC coefficients

VIII c)

In general, three basic redundancies exist in digital images that follow.

1) Psycho-visual Redundancy: The Psychovisual redundancies exist because human perception does not

. I N
involve quantitative analysis of every pixel or luminance value in the image. It corresponds to different

E S
NOT
sensitivities to all image signals by human eyes. Therefore, eliminating some less relative important
information in our visual processing may be acceptable.

KTU
2) Inter-pixel Redundancy: It is a redundancy corresponding to statistical dependencies among pixels,
especially between neighboring pixels. That is, the neighboring pixels are not statistically independent.
The gray levels are not equally probable. The value of any given pixel can be predicated from the value
of its neighbors that is they are highly correlated. The information carried by individual pixel is relatively
small. To reduce the interpixel redundancy the difference between adjacent pixels can be used to
represent an image.

3)Coding Redundancy: •Coding redundancy is associated with the representation of information. The
uncompressed image usually is coded with each pixel by a fixed length. For example, an image with 256
gray scales is represented by an array of 8-bit integers. If the gray levels of an image are coded in a way
that uses more code symbols than absolutely necessary to represent each gray level then the resulting
image is said to contain coding redundancy .Using some variable length code schemes such as Huffman
coding and arithmetic coding may produce compression.

There are different methods to deal with different kinds of aforementioned redundancies. As a result,
an image compressor often uses a multi-step algorithm to reduce these redundancies.

IX a)

Downloaded from Ktunotes.in

The five stages of compression and decompression are shown in Figs. 1 and2.

E S . I N
NOT
All of the steps shown in the compression diagram are invertible, hence lossless, except for the Quantize

KTU
step. Quantizing refers to a reduction of the precision of the floating point values of the wavelet
transform, which are typically either 32-bit or 64-bit floating point numbers. Quantizer design is the
largest factor effecting compression.

•Efficient decorrelation through multiscale decomposition with filter-banks

• Horizontal, vertical and diagonal details with zero mean and Laplacian-like distributions

• Many of the transform coefficients carry little visual information – susceptible for efficient coding

• Intra-scale and inter-scale dependencies (correlations) to be exploited

Main difference with other transform coders is that there is no sub-image division. The localization is an
inherent property of the wavelet transform

Wavelet selection determines the complexity and the compression performance Different wavelet
families: Daubechies (orthogonal, maximal number or vanishing moments), CDF (biorthogonal, spline,
symmetrical), symmlets, etc. The reconstruction functions (and the corresponding filters) are the most
important.

IX b)

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT

Downloaded from Ktunotes.in

 E S . I N
KTU NOT
IX c)

The general compression algorithms use a statistical model to encode single symbols. Compression is
encoding symbols into bit strings that use fewer bits. Dictionary-based algorithms do not encode single
symbols as variable-length bit strings; they encode variable-length strings of symbols as single tokens.
The tokens form an index into a phrase dictionary. If the tokens are smaller than the phrases they
replace, compression occurs. Dictionary-based compression is easier to understand because it uses a
strategy that programmers are familiar with- using indexes into databases to retrieve information from
large amounts of storage like Telephone numbers, Postal codes etc.

The basic idea is:

Given an input source, we want to

 Identify frequent symbol patterns

 Encode those more efficiently
 Use a default (less efficient) encoding for the rest
 Hopefully, the average bit per symbol gets smaller

In general, dictionary-based techniques works well for highly correlated data (e.g. text), but less
efficient for data with low correlation (e.g. i.i.d. sources)
Downloaded from Ktunotes.in
Two dictionary- based compression techniques called LZ77 and LZ78 have been developed. LZ77 is a
"sliding window" technique in which the dictionary consists of a set of fixed- length phrases found in a
"window" into the previously seen text.

E S . I N
KTU NOT

Downloaded from Ktunotes.in