Detection of Microcalcifications in Mammograms using Statistical Measures Based Region Growing
Download
Detection of Microcalcifications in Mammograms using Statistical Measures Based Region Growing Presentation Transcript:
1.Detection of Microcalcifications in Mammograms using Statistical Measures Based Region-Growing
2.Overview
Objective
Introduction
Methodology
Algorithm
Results and Discussions
Conclusions
3.Objective
A novel technique to detect the microcalcifications in digital mammograms.To accurately segment the microcalcifications in the mammogram image. This approach fixes the boundary of the microcalcifications accurately, which confirms its qualitative performance.
4.Introduction
Digital Image Processing
Processing digital images, using computer algorithms
Medical Image Processing
The production of visual representations of body parts, tissues, or organs, for use in clinical diagnosis; encompasses
X-ray methods
Magnetic resonance imaging
Single-photon-emission and positron-emission tomography
Ultrasound.
Region of Interests
It is a sub-image(s) of the entire image, clustering region for segmentation/ analysis.
5.Breast Cancer
Breast cancer is a disease defined by malignant tumors, microcalcifications.
The size, stage, rate of growth, and other characteristics of a breast cancer determine the kind of treatment.
Rarer types of breast cancer require specialized physical examinations.
Image Segmentation
To cluster pixels of similar intensities into relevant image regions.
Regions correspond to subimages or objects.
6.Methodology
Aim: Identifying the microcalcifications
Input: Digital Mammogram Image (I)
Output: Segmented microcalcifications from the input image.
Phase I: Brightness enhancement of the input image
Phase II: Segmentation on the enhanced image using region-growing
7.Algorithm
1. Read the input image I
2. Consider the entire image of M × N do
i. Read k(x,y) /* k(x,y) ? I */
ii. Compute the mean (Imean) for entire image
iii. Identify the row maximum and row minimum intensity of the image as IRmax and IRmin
iv. if k(x,y) >IRmin and k(x,y) k(x,y) = IRmax
else
k(x,y) =IRmin
3. Stop
8.Algorithm
1. Read the enhanced input image IE
2. For the entire image IE do
i. Read kh(x,y) /* kh(x,y) ? IE */
ii. Find the variance as IEVa
where, n is number pixels in kh(x,y) and is the mean.
The standard deviation of the image as
iii. if kh(x,y)>IEVa
kh(x,y)=k(x,y) /* k(x,y) ? I */
else
kh(x,y)=0
9.Results and Discussion
The algorithm is developed in MATLAB 7.8
The performance of the algorithm presented in this paper is evaluated on more than fifty sample images obtained from Mammographic Image Analysis Society (MIAS) database
Using those regions segmented in the enhanced image, as reference, the respective regions are extracted from the original image.
10.Results and Discussion
Download
Detection of Microcalcifications in Mammograms using Statistical Measures Based Region Growing Presentation Transcript:
1.Detection of Microcalcifications in Mammograms using Statistical Measures Based Region-Growing
2.Overview
Objective
Introduction
Methodology
Algorithm
Results and Discussions
Conclusions
3.Objective
A novel technique to detect the microcalcifications in digital mammograms.To accurately segment the microcalcifications in the mammogram image. This approach fixes the boundary of the microcalcifications accurately, which confirms its qualitative performance.
4.Introduction
Digital Image Processing
Processing digital images, using computer algorithms
Medical Image Processing
The production of visual representations of body parts, tissues, or organs, for use in clinical diagnosis; encompasses
X-ray methods
Magnetic resonance imaging
Single-photon-emission and positron-emission tomography
Ultrasound.
Region of Interests
It is a sub-image(s) of the entire image, clustering region for segmentation/ analysis.
5.Breast Cancer
Breast cancer is a disease defined by malignant tumors, microcalcifications.
The size, stage, rate of growth, and other characteristics of a breast cancer determine the kind of treatment.
Rarer types of breast cancer require specialized physical examinations.
Image Segmentation
To cluster pixels of similar intensities into relevant image regions.
Regions correspond to subimages or objects.
6.Methodology
Aim: Identifying the microcalcifications
Input: Digital Mammogram Image (I)
Output: Segmented microcalcifications from the input image.
Phase I: Brightness enhancement of the input image
Phase II: Segmentation on the enhanced image using region-growing
7.Algorithm
1. Read the input image I
2. Consider the entire image of M × N do
i. Read k(x,y) /* k(x,y) ? I */
ii. Compute the mean (Imean) for entire image
iii. Identify the row maximum and row minimum intensity of the image as IRmax and IRmin
iv. if k(x,y) >IRmin and k(x,y)
else
k(x,y) =IRmin
3. Stop
8.Algorithm
1. Read the enhanced input image IE
2. For the entire image IE do
i. Read kh(x,y) /* kh(x,y) ? IE */
ii. Find the variance as IEVa
where, n is number pixels in kh(x,y) and is the mean.
The standard deviation of the image as
iii. if kh(x,y)>IEVa
kh(x,y)=k(x,y) /* k(x,y) ? I */
else
kh(x,y)=0
9.Results and Discussion
The algorithm is developed in MATLAB 7.8
The performance of the algorithm presented in this paper is evaluated on more than fifty sample images obtained from Mammographic Image Analysis Society (MIAS) database
Using those regions segmented in the enhanced image, as reference, the respective regions are extracted from the original image.
10.Results and Discussion
No comments:
Post a Comment