Department Of Computer Engineering

Lab Practice-VI
Final Year Computer Engineering
Semester-VIII
PART-I 410252 : Elective VI :Image Processing
Subject Code: 410252(B)

Prepared By:
Priyadarshini Patil
(Assistant Professor, Computer Engineering)
Sr.
No. Laboratory Assignments
1 Consider any image with size 1024*1024. Modify the image to the sizes 512*512,
256*256, 128*128, 64*64 and 32*32 using subsampling technique. Create the original
image from all the above subsampled images using resampling technique. Read any
image. Display the histogram, Equalized histogram, and image with equalized
histogram
2 Consider any image with size 1024*1024. Modify the image to the sizes 512*512,
256*256, 128*128, 64*64 and 32*32 using subsampling technique. Create the original
image from all the above subsampled images using resampling technique.
3 Read any image. Display the histogram, Equalized histogram, and image with
equalized histogram
4 Read any image. Display the outputs of contrast stretching, intensity level slicing

5 Compare the results of any three edge detection algorithms on the same image dataset
and do the analysis of the result.
6 Mini project: Implement visual surveillance applications and detect moving objects
using object detection and tracking algorithm Or Implement any medical image
processing application for freely available medical image dataset
Assessment:
Laboratory Assignments:

• The external continuous assessment will be 50 marks.

• The outline of distribution assignments Marks for various aspects /mechanisms towards
Continuous Assessment is as follows:

Sr. No. Distribution of Marks Marks

1 Term Work 50
Total 50
 Department of Computer Engineering Laboratory Practice VI

Practical No. 1
Aim : Consider any image with size 1024*1024. Modify the image to the sizes 512*512,
256*256, 128*128, 64*64 and 32*32 using subsampling technique. Create the original image
from all the above subsampled images using resampling technique.

Objective: To learn the resize/scaling/ interpolation concept in Image processing • Input – RGB
images of size 1024*1024.

Theory:
Image digitization (Sampling and quantization) –

In Digital Image Processing, signals captured from the physical world need to be translated into
digital form by “Digitization” Process. In order to become suitable for digital processing, an image
function f(x,y) must be digitized both spatially and in amplitude. This digitization process involves
two main processes called-

1. Sampling: Digitizing the co-ordinate value is called sampling.

2. Quantization: Digitizing the amplitude value is called quantization

The subsampling was accomplished by deleting the appropriate number of rows and columns from
the original image. For example, the 512*512 image was obtained by deleting every other row and
column from the 1024*1024 image. The 256*256 image was generated by deleting every other
row and column in the 512*512 image, and so on. The number of allowed gray levels was kept at
256. These images show the dimensional proportions between various sampling densities, but their
size differences make it difficult to see the effects resulting from a reduction in the number of
samples. The simplest way to compare these effects is to bring all the subsampled images up to
size 1024*1024 by row and column pixel replication.

Quantization is opposite to sampling. It is done on y axis. When you are quantizing an image, you
are actually dividing a signal into quanta(partitions). On the x axis of the signal, are the co-
ordinate values, and on the y axis, we have amplitudes. So digitizing the amplitudes is known as
Quantization.

4 A.Y. 2022-
23
 Department of Computer Engineering Laboratory Practice VI

Procedure:

1. Read and load the images obtained in the previous step of sizes 512*512, 256*256, 128*128,
64*64, 32*32. (One by one).
2. Set the output size as: 1024*1024.
3. Resize the input image to resample it, using the cv2.resize() function by passing the input image
and output size as parameters.
4. Also, set the parameter ‘interpolation’ to cv2.INTER_LINEAR so that it follows linear
interpolation while resampling. (Other types of interpolation can also be tested).
5. Display the resampled images.

Conclusion: We have successfully modified the image to the sizes 512*512, 256*256, 128*128,
64*64 and 32*32 using subsampling technique and create the original image from all the
subsampled images using resampling technique and also display equalized histogram for the
images.

5 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

Practical No. 2
Aim : Consider any image with size 1024*1024. Modify the image to the sizes 512*512,
256*256, 128*128, 64*64 and 32*32 using subsampling technique. Create the original image
from all the above subsampled images using resampling technique. Read any image. Display the
histogram, Equalized histogram, and image with equalized histogram.

Objective: To learn the resize/scaling/ interpolation concept in Image processing • Input – RGB
images of size 1024*1024.

Theory:
Image digitization (Sampling and quantization) –

In Digital Image Processing, signals captured from the physical world need to be translated into
digital form by “Digitization” Process. In order to become suitable for digital processing, an image
function f(x,y) must be digitized both spatially and in amplitude. This digitization process involves
two main processes called-

3. Sampling: Digitizing the co-ordinate value is called sampling.

4. Quantization: Digitizing the amplitude value is called quantization

The subsampling was accomplished by deleting the appropriate number of rows and columns from
the original image. For example, the 512*512 image was obtained by deleting every other row and
column from the 1024*1024 image. The 256*256 image was generated by deleting every other
row and column in the 512*512 image, and so on. The number of allowed gray levels was kept at
256. These images show the dimensional proportions between various sampling densities, but their
size differences make it difficult to see the effects resulting from a reduction in the number of
samples. The simplest way to compare these effects is to bring all the subsampled images up to
size 1024*1024 by row and column pixel replication.

Quantization is opposite to sampling. It is done on y axis. When you are quantizing an image, you
are actually dividing a signal into quanta(partitions). On the x axis of the signal, are the co-
ordinate values, and on the y axis, we have amplitudes. So digitizing the amplitudes is known as
Quantization.

6 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

Procedure:

5. Read and load the images obtained in the previous step of sizes 512*512, 256*256, 128*128,
64*64, 32*32. (One by one).
6. Set the output size as: 1024*1024.
7. Resize the input image to resample it, using the cv2.resize() function by passing the input image
and output size as parameters.
8. Also, set the parameter ‘interpolation’ to cv2.INTER_LINEAR so that it follows linear
interpolation while resampling. (Other types of interpolation can also be tested).
5. Display the resampled images.

Histogram Equalization –

The histogram technique that is used to enhance the brightness and contrast of an image is histogram
equalization. The goal of histogram equalization is to distribute the gray levels within an image so that every
gray level is equally likely to occur. In other words, histogram equalization takes an image’s histogram and
produces a new image with a histogram that is uniformly distributed. Histogram equalization will increase
the brightness and contrast of a dark and low contrast image, making features observable that were not
visible in the original image. Since histogram equalization distributes an image’s gray levels uniformly
about the range of gray levels, all images will have approximately the same brightness and contrast, hence
allowing images to be compared equally without a bias due to perceived contrast and brightness differences.

Conclusion: We have successfully modified the image to the sizes 512*512, 256*256, 128*128,
64*64 and 32*32 using subsampling technique and create the original image from all the
subsampled images using resampling technique and also display equalized histogram for the
images.

7 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

Practical No. 3
Aim : Read any image. Display the histogram, Equalized histogram, and image with equalized
histogram

Objective: To learn to plot image in the form histogram and apply histogram equalization.

Theory:

What we can explore from Histogram? Why Histogram equalization?

We can deduce a great deal about the appearance of the image from its histogram –

i) In a dark image, the gray levels would be clustered at lower ends ii) In an uniformly
bright image, the gray levels would be clustered at the upper end iii) In a well contrasted image,
the gray levels would be well spred out over much of the image

Applications:

For image enhancement

a) Displaying Histogram of input image:
1. Read and load the input RGB image.
2. Convert the image to grayscale using the cv2.cvtColor() function.
3. Obtain the histogram of the grayscale image using the cv2.calcHist() function. Set the
following parameters:
Input image
Channels: [0] (since the image is grayscale).
Mask: None (since we require the histogram of the entire image and not a specific part).
histSize: [256] (represents the bin count; 256 in case of grayscale image).
range: [0, 256] (the range of values that a pixel can have).
4. Plot the histogram using the matplotlib library.

Histogram :

The histogram of an image is a plot of the number of occurrences of gray levels in the image
against gray level values. Histograms are the basis for numerous spatial domain processing techniques. It
plays an important role in enhancement of perceived brightness and contrast of an image. It specifies the

8 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

number of pixels having each gray level, but gives no hint as to where those pixels are located within the
image. The histogram of an N × M image is defined as the percentage of pixels within the image at a given
gray level:

n for 0 i Gmax
MN h= i
i

where ni is the number of pixels at gray level i,

NM is the total number of pixels within the image and
Gmax is the maximum gray level value of the image.

Histogram Equalization :

The histogram technique that is used to enhance the brightness and contrast of an image is
histogram equalization. The goal of histogram equalization is to distribute the gray levels within
an image so that every gray level is equally likely to occur. In other words, histogram equalization
takes an image’s histogram and produces a new image with a histogram that is uniformly
distributed. Histogram equalization will increase the brightness and contrast of a dark and low
contrast image, making features observable that were not visible in the original image. Since
histogram equalization distributes an image’s gray levels uniformly about the range of gray levels,
all images will have approximately the same brightness and contrast, hence allowing images to be
compared equally without a bias due to perceived contrast and brightness differences.

Procedure :

1. Read and load the grayscale image generated in the previous step.
2. Obtain the equalized histogram of the grayscale image using the cv2.equalizeHist() function.
3. Obtain the corresponding histogram which can be displayed for the equalized histogram using
the cv2.calcHist() function. Set the parameters as set above.
4. Plot the equalized histogram using the matplotlib library.
9 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

5. Display the image corresponding to the equalized histogram using cv2 library.

Conclusion: We have successfully read the image and display the histogram, Equalized histogram,
and image with equalized histogram.

10 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

Practical No. 4
Aim : Read any image. Display the outputs of contrast stretching, intensity level slicing.

Objective : To Study Various Methods for Image Enhancement using Spatial and Frequency
Domain.

Theory:

Image enhancement –

The principal objective of enhancement is to process an image so that the resultis more suitable
than the original image for a specific application. Image enhancement approaches fall into two
broad categories: spatial domain methods and frequency domain methods. The term spatial
domain refers to the image plane itself, and approaches in this category are based on direct
manipulation of pixels in an image. Frequency domain processing techniques are basedon
modifying the Fourier transform of an image.
Image enhancement is the process of adjusting digital images so that the results are more suitable
for display or further image analysis. For example, you can remove noise, sharpen, or brighten an
image, making it easier to identify key features.

Contrast Stretching:

One of the simplest piecewise linear functions (function whose pieces are linear) is a
contrast stretching transformation. The form of the piecewise function can be arbitrary complex. A
practical implementation of some important transformation can be formulated only as piecewise
function. It increases the dynamic range of the gray levels in the image being processed. Low
contrast images are resulted from poor illumination, lack of dynamic range in the imaging sensor,
or even wrong setting of a lens aperture of image acquisition.
Transformation function is shown in the graph. Locations of the points (r1, s1) and (r2, s2)
used to control the shape of the transformation function. If r1=s1 and r2=s2 transformation is
linear function and produces no changes in the gray level. If r1=r2, s1=0 and s2=L-1,
transformation becomes a thresholding function that results a binary image. Intermediate values of
(r1, s1) and (r2, s2) produce various degrees of spread in the gray levels of the output image, thus
affecting its contrast.

11 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

Fig : Contrast Stretching

Intensity level slicing-

Intensity level slicing means highlighting a specific range of intensities in an image. In other
words, we segment certain gray level regions from the rest of the image. Suppose in an image,
your region of interest always take value between say 80 to 150, So, intensity level slicing
highlights this range and now instead of looking at the whole image, one can now focus on the
highlighted region of interest. Since, one can think of it as piecewise linear transformation
function so this can be implemented in several ways. Here, we will discuss the two basic type of
slicing that is more often used.

In the first type, we display the desired range of intensities in white and suppress all other
intensities to black or vice versa. This results in a binary image. The transformation function for
both the cases is shown below.

12 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

In the second type. we brighten or darken the desired range of intensities(a to b as shown below)
and leave other inten- sities unchanged or vice versa. The transformation function for both the
cases, first where the desired range is changed and second where it is unchanged, is shown below.

Procedure:

a) Contrast Stretching -
Applying contrast stretching to the input image:
1. Read and load the input RGB image.
2. Split the R, G, B channels of the image.
3. For each channel:
• Initialise the stretch factors.
13 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

• Determine the min and max values in each channel (max_val, min_val).
• Calculate the stretch factor for each channel using the following formula:
• stretch_factor_1 = (stretch_max - stretch_min) / (max_val - min_val)
• stretch_factor_2 = stretch_min - stretch_factor_1 * min_val

4. Initialize the stretched image with zeroes.

5. Obtain the three stretched channels one by one using cv2.convertScaleAbs() function by passing
the above calculated stretch factors to it.
6. Store the stretched channels as parts of the stretched image.
 7. Display the contrast stretched image using cv2 library.

b) Intensity level Slicing

Applying intensity level slicing to the input image:
1. 1.Read and load the input RGB image.
2. Convert the image to grayscale using the cv2.cvtColor() function.
3. Set the lower and upper intensity levels.
4. Create a mask with the intensity range using the intensity levels set above and the obtained
grayscale image. 5. Apply this mask to the original image by performing a bitwise AND to
obtain the intensity sliced image.
 6. Display the intensity sliced image using the cv2 library.

Conclusion: We have successfully read the image and display the outputs of contrast stretching,
intensity level slicing.

14 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

Practical No. 5
Aim : Compare the results of any three edge detection algorithms on the same image dataset and
do the analysis of the result.

Objective : To Learn Classification Techniques for Image Segmentation.

Theory:

Image Segmentation
The first step in image analysis is to segment the image. Segmentation is the process to
subdivide the image into its constituent parts or objects. The level to which subdivision is carried
depends on the problem being solved.
In computer vision, image segmentation is the process of partitioning a digital image into
multiple segments (sets of pixels, also known as super-pixels). The goal of segmentation is to
simplify and/or change the representation of an image into something that is more meaningful and
easier to analyze. Image segmentation is typically used to locate objects and boundaries (lines,
curves, etc.) in images. More precisely, image segmentation is the process of assigning a label to
every pixel in an image such that pixels with the same label share certain characteristics.
The result of image segmentation is a set of segments that collectively cover the entire image, or
a set of contours extracted from the image (see edge detection). Each of the pixels in a region are
similar with respect to some characteristic or computed property, such as color, intensity, or
texture. Adjacent regions are significantly different with respect to the same characteristic(s).When
applied to a stack of images, typical in medical imaging, the resulting contours after image
segmentation can be used to create 3D reconstructions with the help of interpolation algorithms
like Marching cubes.

Segmentation algorithms for monochrome images are generally based on one of the two basic properties
of grey level values:

1. Discontinuity: In this category the approach is to partition the image based on abrupt changes in grey
level.
2. Similarity: In this category the approaches are based in thresholding, region growing, region splitting
and merging.
15 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

Edge Detection :

Edge detection is by far the most common approach for detecting meaningful discontinuities in gray level.
The reason is that isolated points and the thin lines are not frequent occurrences in most practical
applications. An edge is the boundary between two regions with relative distinct gray level properties.
Note from the profile that an edge (transition from dark to light) is modeled as a smooth, rather than as an
abrupt, change of gray level. This model reflects the fact that edges in digital images are generally slightly
blurred as a result of sampling.

1. Gradient Operator :
The gradient of an image f (x,y) at the location (x,y) is defined as

Gradient vector always points in the direction of maximum rate of change of f at co-ordinates (x,y).
Magnitude of the vector is the main quantity in edge detection and given by,

Different Gradient Operators:

Use of different gradient operators for the following 3x3 grey level image is discussed below.
Z1 Z2 Z3
Z4 Z5 Z6
Z7 Z8 Z9
3x3 Grey Level Image
2. Roberts Operator:
One of the simplest way to implement first order partial derivative is to use Roberts cross-gradient
operator.The mask for Roberts operator are given by,

-1 0 0 -1

16 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

0 -1 1 0

The two cross differences for x and y gradient components are given by equation, Gx
= Z9 - Z5 and Gy = Z6 - Z8.
Then we can implement first order derivative ▼f at point Z5.
Mask of size 2x2 is awkward to implement because they don’t have clear center. So generally, 3x3
masks are used.

Sobel operator edge detection algorithm – 1.

Read and load the input RGB image.
2. Convert the image to grayscale using the cv2.cvtColor() function.
3. Apply the Sobel operator to find the x and y gradients of the image using the cv2.Sobel()
function.
4. Calculate the magnitude and angle of the gradients using the following formulae:
• magnitude = np.sqrt(sobel_x**2 + sobel_y**2)
• angle = np.arctan2(sobel_y, sobel_x)
1. Normalize the magnitude to the range [0, 255].
2. Create a binary image based on the gradient magnitudes.
3. Display the binary image with the detected edges using the cv2 library.

Canny edge detection algorithm –

1. Read and load the input RGB image.
2. Convert the image to grayscale using the cv2.cvtColor() function.
3. Apply Gaussian blur to reduce noise.
4. Use the Canny edge detector to find the edges by setting the upper and lower thresholds and
using the cv2.Canny() function.
5. Create a binary image using the obtained edges.

17 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice VI

6. Display the binary image with the detected edges using the cv2 library.

Laplacian operator edge detection algorithm –

1. Read and load the input RGB image.
2. Convert the image to grayscale using the cv2.cvtColor() function.
3. Apply Gaussian blur to reduce noise.
4. Use the Laplacian operator edge detector to find the edges using the cv2.Laplacian()
function.
5. Normalize the magnitude to the range [0, 255].
6. Create a binary image using the obtained edges.
7. Display the binary image with the detected edges using the cv2 library.

Conclusion : We have successfully compare the results of edge detection algorithms on the image dataset
and analysis of the result.

18 A.Y. 2022-23
 Department of Computer Engineering Laboratory Practice
VI

Practical No. 6

Mini Project

Mini project 1 –

Implement visual surveillance applications and detect moving objects using object detection
and tracking algorithm.

Mini project 2–

Implement any medical image processing application for freely available medical image
dataset.

Mini project 3 –

Implement image segmentation to detect object in the background of image.

19 A.Y. 2022-
23