ECE433 COMPUTER VISION PROJECT
DIGITAL IMAGES
PROCESSING
Presented by: Nguyễn Xuân Đạt
Liêu Phong
Phan Anh Minh
Trịnh Trường Vũ
Hồ Trung Tới
OVERVIEW
DIGITAL IMAGES PRACTICAL
INTRODUCTION PROCESSING METHODS APPLICATION
Introduction about DIP Introduction methods Introduction practical
methods and demo code application
INTRODUCTION
What is the definition of “Digital Images Processing” ?
PATICIPANTS
Introduction
What is DIP ?
Digital image processing is using
computer algorithms to perform
various operations on digital
images, to enhance, manipulate, or
extract useful information from the
image. It involves the application
of mathematical and
computational techniques to
digitized image data.
DIGITAL IMAGES
PROCESSING METHODS
DIGITAL IMAGES
PROCESSING lo g ical
o
Histogra Morph
m Gray Level e ra tion
op
Transformations
METHODS
There are numerous digital image
processing techniques available,
such as image segmentation, edge
detection, and color correction;
however, for this project, we will
primarily focus on the utilization of Image
segmen
tation Spatial
the following methods.
filters
HISTOGRAM
Introduction
What is HISTOGRAM ?
A histogram is a graphical
representation of the distribution
of pixel values in an image. It
shows the frequency or count of
each pixel value (typically
ranging from 0 to 255 for an 8-
bit image) within the image.
Introduction
What is Histogram
Equalization ?
Spreading out the frequencies in an image (or equalizing the
image) is a simple way to improve dark or washed-out images.
The formula for histogram equalization is given where:
rk: input intensity
SK: processed intensity
k: the intensity range (e.g 0 – 255)
nj: the frequency of intensity j
M.N: the total number of pixels
DEMO
BEFORE
AFTER
GRAY LEVEL
TRANSFORMATIONS
Introduction
What is LOG operation ?
Log transformations are an image
processing technique that calculates
the logarithm of pixel values to adjust
brightness and contrast, helping to
highlight details in dark areas and
reduce glare in bright areas, thus
improving image quality.
DEMO
BEFORE
AFTER
MORPHOLOGICAL
OPERATION
Introduction
What is AND logical
operation?
The AND operation is a binary operation that
compares corresponding pixels between two input
images or an image and a binary value, producing an
output image where only the pixels that match the
binary condition are retained.
DEMO
BEFORE
AFTER
BOUNDARY OPERATOR
First, erode A by B, then make a set difference between A
and the erosion.
The thickness of the contour depends on the size of the
constructed object – B.
DEMO
BEFORE AFTER
Introduction
What is Hit-or-Miss
Transformation?
The hit-and-miss transform is a binary
morphological operation that searches
for specific patterns of foreground and
background pixels within an image.
Introduction
What is Hit-or-Miss
Transformation?
Four structuring elements are used for corner finding in
binary images using the hit-and-miss transform.
Note that they are all the same element, but rotated by
different amounts.
DEMO
ORIGIN RESULT BINARY
Introduction
What is Opening
method?
Opening is a morphological operation
that removes small, unwanted objects or
noise from an image while preserving the
overall shape and size of larger, more
important objects.
Introduction
How does the Opening method work?
1. Erosion: using a structuring element to shrink or remove small
objects and details, leaving larger objects relatively unchanged.
2. Dilation: using the same structuring element to restore the size
of the remaining larger objects, keeping the small objects and
details removed.
DEMO
ORIGIN RESULT
Introduction
What is “Label Extraction
method”?
Identify and distinguish different objects,
regions, or features within an image.
The process of labeling involves assigning
a unique identifier or label to each distinct
object or region in the image.
Identifier example:
t : red
e: green
DEMO
ORIGIN RESULT
IMAGE SEGMENTATION
EDGE DETECTION METHODS
What is the techniques of edge detection and their
pros and cons ?
Introduction
What is Edge?
In an image, an edge represents a
boundary or contour where there is a
significant change in certain physical
aspects of the image (such as
brightness, surface reflectance, etc.).
Introduction
What is Edge Detection?
Edge detection is a technique in image
processing and computer vision used to
identify and highlight boundaries or
edges within an image.
EDGE DETECTION
METHODS
Basing on 2 main operators
Gradient Based : based operator which
computes first-order derivations in a digital
image like, Sobel operator, Prewitt operator,
Robert operator
Gaussian Based : based operator which
computes second-order derivations in a
digital image like, Canny edge detector,
Laplacian of Gaussian
Prewitt Operator
This operator is almost similar to
the Sobel operator. It also detects
the horizontal and -45 degrees
edges of an image. It is one of the
best ways to detect the
orientation and magnitude of an
image. It uses the kernels or
masks
DEMO
Demo edge detection - Prewitt
BEFORE AFTER
PREWITT OPERATOR
Advantages Disadvantages
01 Good performance on detecting
vertical and horizontal edges 01 The magnitude of coefficient is
fixed and cannot be changed
02 02
Best operator to detect the Diagonal direction points are
orientation of an image not preserved always
Introduction
What is Thresholding?
image segmentation technique that partitions or
separates an image into multiple segments or regions
based on the pixel intensity values.
Introduction
How Thresholding works
Identify a suitable intensity value, called the
threshold
Divide the pixels in an image into two or more
classes or segments.
+ Pixels with values above the threshold are
considered to belong to one class
+ Pixels with values below the threshold are
considered to belong to another class
DEMO
Demo Thresholding
BEFORE AFTER
Introduction
What is Otsu?
Image segmentation technique that automatically
performs histogram shape-based image thresholding,
which is the process of partitioning an image into
foreground and background.
Introduction
How Otsu works
Compute the histogram and probabilities of each
intensity level.
Set an initial threshold t than compute:
Select the threshold value t that maximizes the
between-class variance or minimizes the within-class
variance.
DEMO
Demo Otsu
BEFORE AFTER
THRESHOLDING OTSU
SPATIAL FILTERS
Introduction
What is Median filter?
A median filter is a nonlinear
digital filtering technique that
reduces noise while preserving
edges. Used to remove salt-
and-pepper noise.
Introduction
How does a Median filter work?
Replace each pixel value in an image with the median of the
pixel values in a surrounding neighborhood.
The neighborhood is typically a square region centered on the
pixel being processed.
DEMO
NOISE IMAGE RESULT
Introduction
What is What is Image averaging?
Image averaging is a technique where multiple images
of the same scene or object are combined to create a
single, improved image.
Introduction
How does a Median filter work?
The general implementation for filtering an MXN image with a
weighted averaging filter of size mxn is given by
where a=(m-1)/2 and b=(n-1)/2
DEMO
NOISE IMAGE RESULT
THANKS FOR
LISTENNING
