Image Enhancement - PowerPoint PPT Presentation

Loading...

PPT – Image Enhancement PowerPoint presentation | free to download - id: 5585c-ZDBiY



Loading


The Adobe Flash plugin is needed to view this content

Get the plugin now

View by Category
About This Presentation
Title:

Image Enhancement

Description:

... and Woods, ' Digital Image Processing,' 2nd Edition, Prentice ... The display will tend to produce images darker than intended. So the display is distorted. ... – PowerPoint PPT presentation

Number of Views:108
Avg rating:3.0/5.0
Slides: 28
Provided by: aso9
Learn more at: http://cs.gmu.edu
Category:

less

Write a Comment
User Comments (0)
Transcript and Presenter's Notes

Title: Image Enhancement


1
Image Enhancement
2
References
  • Gonzalez and Woods, Digital Image Processing,
    2nd Edition, Prentice Hall, 2002.
  • Jahne, Digital Image Processing, 5th Edition,
    Springer 2002.
  • Jain, Fundamentals of Digital Image Processing,
    Prentice Hall 1989

3
Overview
  • Human perception (focus of this discussion)
  • Machine perception (ocr)
  • Application specific
  • Heuristic based result better than the original
    image subjective assessment
  • Spatial vs frequency domain

4
Spatial Domain
  • Based on the collection of pixels in the image
  • Enhancement techniques will yield
  • Noise reduction
  • Neighborhood smoothing
  • Highlighting of desired feature

5
Spatial Domain Math Framework
From 1
Typically, spatial domain based enhancement
involves g(x,y) T f(x,y), where f input
image g output image T operator defined on
f based on a neighborhood of x,y. If the
neighborhood is 1x1 pixel, then the output
intensity is dependent on the current intensity
value of the pixel, and can be represented as s
T( r )where r and s are gray level values of
f(x,y) and g(x,y) at location x,y. In such
situations T is a gray-level transformation
function.
6
Examples of Gray level Transformation Functions
(Point Processing)
From 1
Contrast Stretching Best if input is 0 outside a
range of values.
Thresholding Result is a binary Image
7
Larger Neighborhoods
  • Objective determine g(x,y) based on input
    intensity (gray level) values f(x,y) in the
    neighborhood of x,y.
  • Mask processing or filtering
  • Each of the elements in the neighborhood has an
    associated weight
  • g(x,y) depends on f(a,b)a,b ? N(x,y)

8
Basic Gray Level Transformations
Light
From 1
Dark
Dark
Light
9
Image Negatives
From 1
In this example, using the image negative, it is
easier to analyze the breast tissue.
10
Log transformations
  • s c log (1 r)
  • r gt 0 hence 1 r gt 0 log 0 ?
  • Log transformations are useful, when there is a
    large dynamic range for the input variable ( r ).

From 1
Range 0 to 1.5106
Range 0 to 6.2
11
Power law transformation
From 1
Stretch higher (lighter) gray levels
Many display devices (e.g. CRT) respond like the
power law, i.e intensity voltage relationship
is power law based gamma 1.8 to 2.5. The display
will tend to produce images darker than intended.
So the display is distorted. Gamma correction
is used to correct for this distortion.
Stretch lower (ldarker) gray levels
12
Display distortion correction
  • Gamma correction can also fix the distortions in
    color.
  • More important with the internet.
  • Many viewers, variety of monitors.
  • Gamma of view station is not known.
  • Preprocess using an average gamma.

From 1
13
Power law contrast manipulation
(c) better than (b). (d) Background is better
than (c) but washed out effect.
From 1
14
Piece-wise Linear Transformation Contrast
Stretching
From 1
Mean gray level value of the image
15
Gray level Slicing
From 1
16
Bit plane Slicing
MSb (bit 7)
From 1
LSb (bit 0)
17
Histograms
  • Histogram - frequency of occurrence of a gray
    level value
  • Normalizing histograms with rest to the total
    number of pixels converts these into probability
    density like function
  • Histogram processing yields robust image
    processing results
  • Histograms are NOT unique

18
Histograms for 4 images
  • For high contrast, it is best to have a larger
    range of gray level values.
  • If we could transform an image with a resulting
    change in histogram, then that may yield more
    contrast.
  • We need to study the rules for transforming
    histograms, and study the resulting impact on
    images.

From 1
19
Transformation
From 1
s
  • Applying transformations to histograms, can use
    results from probability.
  • Consider the transformation S T( r ) 0 lt r
    lt1 which has the following properties
  • T( r ) is single-valued and monotonically
    increasing in 0,1
  • 0 lt T( r ) lt 1 for r in 0,1
  • Single valued condition ensures that an inverse
    transformation exists, and the monotonic
    condition ensures the increasing order of the
    pixels from black to white.
  • (a) and (b) do not ensure that the inverse
    transform is single valued.

s1
r
r
s1
s
20
Histogram Equalization
From 1
21
Transformation Functions
From 1
22
Mapping for Histogram Specification
From 1
23
Example of Histogram Specification
From 1
24
Continued
From 1
25
Localized Histogram Equalization
From 1
26
From 1
27
Histogram Stats
From 1
About PowerShow.com