Scribd
Upload
61 views30 pages

Image Segmentation Techniques Explained

Uploaded by

DK drift king
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd

Topics covered

  • Image Attributes,
  • Point Detection,
  • Intensity Similarity,
  • Background Pixels,
  • Line Detection,
  • Image Segmentation Application…,
  • Adaptive Techniques,
  • Image Histogram,
  • Segmentation Challenges,
  • Intensity Discontinuity
61 views30 pages

Image Segmentation Techniques Explained

Uploaded by

DK drift king
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPT, PDF, TXT or read online on Scribd

Topics covered

  • Image Attributes,
  • Point Detection,
  • Intensity Similarity,
  • Background Pixels,
  • Line Detection,
  • Image Segmentation Application…,
  • Adaptive Techniques,
  • Image Histogram,
  • Segmentation Challenges,
  • Intensity Discontinuity

Digital Image

Processing

Image Segmentation
Dr. Fawad Nasim
[Link]@[Link]
Contents

• What is Segmentation?
• What is thresholding?
• Simple thresholding
• Adaptive thresholding
Image segmentation
Image segmentation is a computer vision technique that partitions a digital image
into discrete groups of pixels—image segments—to inform object detection and
related tasks.

Each segment:
Looks uniform
Belongs to a single object.
Have some uniform attributes.
All the pixels related to it are connected.
Detection of Discontinuities

 Point Detection
 Line Detection

 Edge Detection
Detection of Discontinuities

Matrix notation
w = [w1 w2 ... wn] maske
Z = [ z1 z2 … zn] tilhørende billed pixels
R(x,y) = [w] [Z]’

Convolution with 3x3 mask


R w1 z1  w2 z 2  .....  w9 z9
i 9 i n
 wi zi or general  wi zi
i 1 i 1
Point detection

R T
T is a non-zero threshold
Segmentation Algorithms
 Segmentation algorithms are based on one of
two basic properties of intensity values
discontinuity and similarity.
 First category is to partition an image based on

abrupt changes in intensity, such as edges in


an image.
 Second category are based on partitioning an

image into regions that are similar according to


a predefined criteria. Thresholding, Histogram
Thresholding approach falls under this
category.
Thresholding

 Thresholding is usually the first


step in any segmentation approach
 We have talked about simple

single value thresholding already


 Single value thresholding can be

1 if f ( x, y ) asT follows:
given mathematically
g ( x, y ) 
0 if f ( x, y ) T
Thresholding

3 5 7 3 4 2 1
2 4 9 10 22 9 3
3 5 12 11 15 10 3
5 6 11 9 17 19 1
2 3 11 12 18 16 2
3 6 8 10 18 9 5
4 6 7 8 3 3 1
Thresholding

3 5 7 3 4 2 1
2 4 9 10 22 9 3
3 5 12 11 15 10 3
5 6 11 9 17 19 1
2 3 11 12 18 16 2
3 6 8 10 18 9 5
4 6 7 8 3 3 1
Thresholding

3 5 7 3 4 2 1
2 4 9 10 22 9 3
3 5 12 11 15 10 3
5 6 11 9 17 19 1
2 3 11 12 18 16 2
3 6 8 10 18 9 5
4 6 7 8 3 3 1
Thresholding

3 5 7 3 4 2 1
2 4 9 10 22 9 3
3 5 12 11 15 10 3
5 6 11 9 17 19 1
2 3 11 12 18 16 2
3 6 8 10 18 9 5
4 6 7 8 3 3 1
Example

A 4x4 grayscale image with the following pixel intensities:


Example

Step 1: Calculate the Average Intensity


•Sum of all pixel intensities: 1045
•Total number of pixels: 16
•Average intensity (T): 1045 / 16 ≈ 65.31

Step 2: Initial Thresholding


•Set the initial threshold to 65.
•Compare each pixel value to the threshold:
• If the pixel value is greater than 65,
assign it to the foreground (e.g., 1).
• If the pixel value is less than or equal
to 65, assign it to the background (e.g.,
0).
Example

Step 3: Update Threshold


•Calculate the average intensity of the
foreground and background pixels:
• Average intensity of foreground pixels:
(104 + 113 + 126) / 3 ≈ 111
• Average intensity of background
pixels: (52 + 55 + ... + 64) / 13 ≈ 58.54
•New threshold: (111 + 58.54) / 2 ≈ 84.77

Second Iteration:
•Reclassify pixels based on the new threshold of 84.77.
Note: This is a simplified example. In practice, iterative thresholding can be
continued until the threshold converges or a maximum number of iterations is
reached. Additionally, more sophisticated techniques like Otsu's method can
be used to automatically determine the optimal threshold value.
Thresholding Example

 Imagine a poker playing robot that


needs to visually interpret the
cards in its hand

Original Image Thresholded Image


But Be Careful

 Ifyou get the threshold wrong the


results can be disastrous

Threshold Too Low Threshold Too High


Basic Global Thresholding
 Based on the histogram of an image
 Partition the image histogram using a

single global threshold


 The success of this technique very

strongly depends on how well the


histogram can be partitioned
Basic Global Thresholding Algorithm
 The basic global threshold, T, is calculated
as follows:
1. Select an initial estimate for T (typically the
average grey level in the image)
2. Segment the image using T to produce two
groups of pixels: G1 consisting of pixels with
grey levels >T and G2 consisting pixels with
grey levels ≤ T
3. Compute the average grey levels of pixels in
G1 to give μ1 and G2 to give μ2
Basic Global Thresholding Algorithm

4. Compute a new threshold value:


1   2
T
2
5. Repeat steps 2 – 4 until the difference in
T in successive iterations is less than a
predefined limit T∞
This algorithm works very well for
finding thresholds when the histogram
is suitable
Thresholding Example 1
Thresholding Example 2
Problems With Single Value
Thresholding
 Single value thresholding only
works for bimodal histograms
 Images with other kinds of

histograms need more than a


single threshold
Problems With Single Value
• Let’s say weThresholding
want to (cont…)
isolate the contents
of the bottles

• Think about what the


histogram for this
image would look like

• What would happen if we used a single


threshold value?
Single Value Thresholding and
Illumination

 Uneven illumination can really upset a


single valued thresholding scheme
Basic Adaptive Thresholding

 An approach to handling situations in


which single value thresholding will
not work is to divide an image into
sub images and threshold these
individually
 Since the threshold for each pixel

depends on its location within an


image this technique is said to
adaptive
Basic Adaptive Thresholding Example
 Theimage below shows an example of
using adaptive thresholding with the
image shown previously

 As can be seen success is mixed


 But, we can further subdivide the

troublesome sub images for more


success
Basic Adaptive Thresholding Example
(cont…)
 These images show the
troublesome parts of the
previous problem further
subdivided
 After this sub division

successful thresholding
can be
achieved
Segmentation Based on Clustering

Common questions

Powered by AI

The basic global thresholding algorithm begins by selecting an initial threshold, typically the average grey level. It segments the image into two pixel groups, computes their average grey levels, and adjusts the threshold iteratively until changes fall below a set limit . Its limitation includes reliance on a suitable histogram distribution, making it ineffective for non-bimodal histograms .

Single value thresholding fails in scenarios of uneven illumination or when the image histogram is not bimodal. Adaptive thresholding addresses this by varying the threshold across image regions, thus better accommodating local intensity variations and ensuring accurate segmentation .

Iterative thresholding involves calculating an average initial threshold and reclassifying pixel values accordingly. For a 4x4 image, an initial threshold might classify pixels into foreground and background based on intensity. A recalculated threshold, considering the average of these categories, refines segmentation progressively until it stabilizes, enhancing classification accuracy .

Variations in illumination can disrupt single value thresholding, leading to misclassification of pixel intensities. Mitigation strategies include using adaptive thresholding, which accounts for local variations, thereby ensuring consistent segmentation despite uneven lighting conditions .

An incorrect threshold can misclassify pixel regions, leading to either loss of crucial details or excessive noise in the segmented image. Such errors adversely impact object detection and analysis tasks, as demonstrated in scenarios where automated interpretation requires precise segmentation .

The success of global thresholding is contingent on the ability to partition the image histogram into distinct groups. In bimodal histograms, where foreground and background are distinctly separated, thresholding is effective, while unsuitable histograms, with overlapping intensities, lead to poor segmentation results .

Discontinuity detection identifies changes like edges within an image, crucial for defining object boundaries. However, its effectiveness is limited when dealing with textures or low contrast, necessitating complementary methods to achieve comprehensive image segmentation .

Segmentation algorithms are influenced by intensity properties such as discontinuity for edge detection and similarity for region partitioning. Effectiveness depends on the algorithm's ability to accurately identify and process these properties; for example, edge detection is suited for images with clear intensity changes, while region-based methods work well with homogenous regions .

Image segmentation algorithms are classified based on two main properties: intensity value discontinuity and similarity. Algorithms focusing on discontinuity detect abrupt changes in intensity, such as edges, while those based on similarity partition images into regions according to predefined criteria .

Adaptive thresholding divides an image into smaller segments, applying a local threshold to each, thus accommodating varying illumination and intensity levels across the image. This method successfully addresses situations where single value thresholding fails due to uneven illumination .

You might also like