Software Engineering Final Project: Image Processing Use Cases - Prof. Xiaojun Qi, Assignments of Software Engineering

Download Software Engineering Final Project: Image Processing Use Cases - Prof. Xiaojun Qi and more Assignments Software Engineering in PDF only on Docsity!

CS2450 SOFTWARE ENGINEERING

FINAL PROJECTFINAL PROJECT

FALL 2006

TEAM:TEAM:

Michael XXX – Distribution 1

Christopher XXX – Distribution 2

Troy XXX – Distribution 3

Digital Image Processor^ l^

l

Color Image to Grayscale Image

(5 methods) Enhance Image(3 methods)(^

Noise Removal from Image

User/Qi

Noise

Removal from Image

(2 methods)

Edge Detection in Image

(1 method)

Brief Description:

Enhance Image

The

Enhance Image

use case enables the user/Qi to enhance a

grayscale image. Step by Step DescriptionStep-by-Step Description 1. User opens an image 2

Di

l^

filt

d i

2. Display unfiltered image*3. User chooses

Enhance

method from 3 possible methods:

•^

Histogram equalization

-^

Scaling

-^

Power transformation

4. Convert image to grayscale

g^

g^

y

5. Convert image / Run Enhance method6. Display modified image* 7

Save image

7. Save image*if display is active (GUI only)

Brief Description:

Noise Removal from Image

p^

f^

g

The

Noise Removal from Image

use case enables the user/Qi to

remove noise from a grayscale image. Step-by-Step Description 1

User opens an image

1. User opens an image2. Display unfiltered image*3. User chooses

Noise Removal

method from 2 possible methods:

•^

Mean Filter

-^

Mean

Filter

•^

Median Filter

4. Convert image to grayscale 5

C

t i

/ R

N i

R

l^

th d

5. Convert image / Run Noise Removal method6. Display modified image7. Save imageif display is active (GUI only)

A

nalysis Phase

Color to GrayscaleColor

to Grayscale

C

l^

I^

t^

G

l^

I^

USE CASE

C

olor Image to Grayscale Image

- USE CASE

SCENARIO (NORMAL)

-^

User opens imageUser opens image

-^

System displays unfiltered image

-^

User selects

Color Image to Grayscale Image

User selects

Color

Image to Grayscale Image

-^

User selects method:

Color Vector, Simple Average,

,^

p

g ,

Weighted Average, NTSC/PAL, ITU-R

-^

System retrieves image format information

-^

System filters image

-^

System displays image

Image EnhancementImage

Enhancement

E

h

I^

USE CASE SCENARIO (NORMAL)

E

nhance Image

- USE CASE SCENARIO (NORMAL) -^

User opens grayscale image

-^

System displays unfiltered image

-^

System displays unfiltered image

-^

User selects

Enhance Image

-^

User selects method:

-^

User

selects method: Histogram Equalization, Scaling,Power TransformationPower Transformation

-^

System retrieves image format information

-^

System filters image

y^

g

-^

System displays image

Edge DetectionEdge

Detection

Edge Detection in Image

- USE CASE

SCENARIO (NORMAL)

•^

User opens grayscale image

•^

System displays unfiltered imageSystem displays unfiltered image

•^

User selects

Edge Detection in Image

(Canny

Edge Detector only

Edge Detector only…)

•^

System retrieves image format information

•^

System filters image

•^

Systems displays image

Part 2: Entity Class ModelingPart

2: Entity Class Modeling

• Program description and noun extraction:

DIMP allows a user to manipulate digital images by changing colorDIMP allows a user to manipulate digital images by changing colorto grayscale, enhancing grayscale image, remove noise from agrayscale image, and detect edge in a grayscale image.

Initial Class Diagram:

ImageWidthH i hHeightFormat

Part 4: Extract Boundary and

Control Classes

• Boundary Class:

User Interface ClassUser Interface Class

C

t^

l Cl

• Control Classes:

Grayscale

y

NoiseEnhanceEdge

Part 5: Use Case RealizationPart

5: Use Case Realization

Collaboration diagram for Enhance use case

Code OrganizationCode

Organization

• Modular Code:

Each of our separate parts was modeledEach of our separate parts was modeledas a class.

Mi h

l C l

t^

G

l^

C

i

– Michael: Color to Grayscale Conversion– Chris: Image Enhancements– Troy: Noise Filter

• The Canny edge detector was also

modeled as a classmodeled as a class

• Each of the six parts made into separate

th d

methods– Troy: Gaussian Filter and Edge Strength

y

g

g

• Chris: Edge Direction and Edge Normalization– Michael: Non maximum Suppression and– Michael: Non maximum Suppression and

Hysteresis