EE-333 Digital Image Processing

Image Enhancement in Spatial Domain

Dr Tahir Nawaz
Website: https://www.tahirnawaz.com
Email: [email protected]
Image enhancement

Original image Enhanced image

Image enhancement

Original image Enhanced image

Image enhancement
• Processing an image so that the result is more suitable
and improved than the original image for a specific
application.

• Image Enhancement Methods

– Spatial Domain: Direct manipulation of pixels in an
image
– Frequency Domain: Process the image by modifying
the Fourier transform of an image

• This Lecture – Spatial Domain

Types of image enhancement operations
• Point/Pixel operations
– Output value at specific
coordinates (x,y) is
dependent only on the input
value at (x,y)
• Local operations
– The output value at (x,y) is
dependent on the input
values in the neighborhood of
(x,y)
• Global operations
– The output value at (x,y) is
dependent on all the values
in the input image
Processing in spatial domain
• Most spatial domain enhancement operations can be
generalized as:

g(x,y) = T [f(x,y)]

• f(x, y) = the input image

• g(x, y) = the processed/output image

• T = an operator defined over some neighborhood of (x,y)

Processing in spatial domain
Intensity or mapping transformation
• The most simplest form of T is when the neighborhood is
1x1, i.e. single pixel.
• s = T(r)
– For simplicity, r and s are variables denoting gray level of f(x,y)
and g(x,y), respectively, at any point (x,y)
Intensity or mapping transformation
• Segmentation of an object of interest from a background
Intensity or mapping transformation
• Intensity scaling
s = T(r) = a.r
Basic gray-level transformations
• Some basic intensity transformation functions. All curves
are scaled to fit in the range shown.
Basic gray-level transformations
• Image Negatives
– Reverses the gray level order
– For L gray levels, the transformation has the form:
s = (L - 1) - r
– Negative images are useful for enhancing white or grey detail
embedded in dark regions of an image
Basic gray-level transformations
• Log Transformations
– The general form of the log
transformation is
s = c log (1 + r)
– The log transformation maps a
narrow(wider) range of
low(high) input grey level
values into a wider(narrower)
range of output values
– For lower amplitudes of input
image the range of gray levels
is expanded
– For higher amplitudes of input
image the range of gray levels
is compressed
– The inverse log transformation performs the opposite
transformation
Basic gray-level transformations
• Log Transformations
– This transformation is suitable for the case when the dynamic
range of a processed image far exceeds the capability of the
display device (e.g. display of the Fourier spectrum of an image)
– It is also called “dynamic-range compression/expansion”
Basic gray-level transformations
• Power-Law Transformations
– Power law transformations
have the following form
s=crƔ
– Map a narrow range of dark
input values into a wider range
of output values or vice versa
– Varying Ɣ gives a whole family
of curves

• For Ɣ < 1: Expands values of dark pixels, compress

values of brighter pixels
• For Ɣ > 1: Compresses values of dark pixels, expand
values of brighter pixels
• If Ɣ = 1 & c = 1: Identity transformation (s = r)
Basic gray-level transformations
• Power-Law Transformations
– Result of contrast enhancement
Basic gray-level transformations
• Power-Law Transformations
– Reducing Ɣ too much leads to a wash-out appearance in an
image
Basic gray-level transformations
• Power-Law Transformations
– Image enhancement
Basic gray-level transformations
• Piecewise-Linear Transformations
– A complementary approach to the methods discussed in the
previous three transformations is to use piecewise linear
functions

– Contrast stretching

– Intensity-level slicing
Basic gray-level transformations
• Piecewise-Linear
Transformations
– Contrast stretching
• Increase the dynamic
range of the gray levels
for low contrast images
• Rather than using a well
defined mathematical
function, we divide it
into piecewise linear
segments
Basic gray-level transformations
• Piecewise-Linear Transformations
– Contrast stretching
Basic gray-level transformations
• Piecewise-Linear Transformations
– Gray-level slicing
• It involves highlighting or enhancing a specific range of gray levels
Basic gray-level transformations
• Piecewise-Linear Transformations
– Gray-level slicing
Histogram processing
• Let I be a grayscale image

• I(r,c) stores an 8-bit integer (intensity value) between 0

and 255

• Histogram, hI, of I:
– a 256-element array, hI
– hI (g) = number of pixels in I that have the value g
– where g = 0,1, 2, 3, …, 255
Histogram processing
• More formally, we can define histogram in the form of a
discrete function as follows:
h(rk)=nk,
where rk is the kth gray level value and nk is the number
of pixels having gray level value rk in the image
Histogram processing
• Actually we can have exactly the same histogram of two
different images too!
Histogram processing
• Histogram of a grayscale image
Histogram processing
• Histogram of a grayscale image
Histogram processing
• Histogram of a colored image
– In the case of colored image, I has got three channels

– I(r,c,R), I(r,c,G), and I(r,c,B) are integer values between

0 and 255.

– I actually has 3 histograms:

• hR(g)= # of pixels in I(:,:,1) with intensity value g
• hG(g) = # of pixels in I(:,:,2) with intensity value g
• hB(g) = # of pixels in I(:,:,3) with intensity value g
Histogram processing
• Histogram of a colored image
Histogram processing
• Examples
Histogram processing
• Examples
Histogram processing
• Examples
Histogram processing
• Examples
Histogram equalization
• Histogram equalization re-assigns the intensity values of
pixels in the input image such that the output image
contains a uniform distribution of intensities
Histogram equalization
Histogram equalization
• Aerial photograph of Pentagon
Histogram equalization
• Probability Distribution Function (pdf) of an image
Histogram equalization
• p(g) is the fraction of pixels in an image that have
intensity value g, or more formally, it is the probability
that a pixel randomly selected from the given image has
intensity value g.

• Whereas the sum of the histogram h(g) over all g from 0

to 255 is equal to the number of pixels in the image, the
sum of p(g) over all g is 1.

• pI(g) is the normalized histogram of the image

Histogram equalization
• The Cumulative Distribution Function (CDF) of an image
– Let ɣ = I(r,c) be the value of a randomly selected pixel from I. Let
g be a specific gray level. The probability that ɣ ≤ g is given by:

– where hI(ɣ) is the histogram of image I.

Histogram equalization
• P(g) is the fraction of pixels in an image that have
intensity values less than or equal to g, or more formally,
it is the probability that a pixel randomly selected from
the given band has an intensity value less than or equal
to g.

• P(g) is the cumulative (or running) sum of p(g) from 0

through g inclusive.

• P(0) = p(0) and P(255) = 1

Histogram equalization
• We employ CDF (PI(g)) to perform histogram
equalization of an image I

• Task: to remap image I so that its histogram is as close

to constant as possible

• We could achieve the above task by using CDF as the

look-up table LUT
Histogram equalization
Histogram equalization
Histogram equalization
Histogram equalization
Histogram equalization
• Implementation
Histogram equalization
• Example
Histogram equalization
• Example
cdf

s = round(255 x 0.719) = 183

183
Histogram equalization
• Example
Histogram equalization
• Example
Histogram equalization
• Example
Histogram equalization
• Example
Acknowledgement/References
• Digital Image Processing”, Rafael C. Gonzalez & Richard E. Woods,
Addison-Wesley, 2002
• Statistical Pattern Recognition: A Review – A.K Jain et al., PAMI (22)
2000
• Pattern Recognition and Analysis Course – A.K. Jain, MSU
• “Pattern Classification” by Duda et al., John Wiley & Sons.
• “Machine Vision: Automated Visual Inspection and Robot Vision”,
David Vernon, Prentice Hall, 1991
• www.eu.aibo.com/
• Advances in Human Computer Interaction, Shane Pinder, InTech,
Austria, October 2008
• https://www.cs.nmt.edu/~ip/lectures.html
• https://web.stanford.edu/class/ee368/handouts.html