Uploaded by khalid enowyli

Endometrial ppt

advertisement
EARLY DIAGNOSIS OF ENDOMETRIUM
CANCER USING IMAGE PROCESSING
TECHNIQUES
Here is where your presentation begins
E-learning Infographics
01
Chapter one
02
Chapter two
03
Chapter three
04
INTRODUCTION
THEORETICAL BACKGROUND AND RELTED WORKS
DIGITAL IMAGE PROCESSING TECHNIQUES
Chapter four
EXPERIMENTATION OF ENDOMETRIUM CANCER
05
Chapter five
06
Chapter six
Results
CONCLUSIONS AND FUTURE WORKS
Chapter
1
INTRODUCTION
INTRODUCTION
•
•
According to GLOBOCAN 2020 estimates of cancer incidence
and mortality, Endometrium cancer is the second leading cause
of mortality in women after Breast cancer
The raw input uterine ultrasound image was first improved
during the preprocessing stage by removing background,
unwanted artifacts, and labels in order to find the clear
shape of the endometrium and detect abnormal shape
Problem Statement
PROBLEM STATEMENT
Aim of the thesis
Aim 1
To improve the medical image denoising base on image processing filters
Aim 2
To apply image segmentation, which plays an important role in the image processing
stagesis
Aim 3
To differentiate between normal and abnormal of endometrium shapes.
Aim 4
To study endomtrium cancer
Aim 5
The proposed techniques to be implemented in MATLAB ® and tested on real case
studies.
METHODOLOGY

MOTIVATION TO T H E THESIS
Overcome
The development of endomtrium
cancer
Task 1
Better Cancer Survival Rates
Facilate eraly detection for
endometrium cancer
Task 2
Provide “second opinion” :
Computerized decision support
systems
Task 3
Fast , reliable and cost effective
Motivation to the
research: Goal
early diagnosis of endometrial cancer is
very important to reduce mortality rate
of
women
developing
a
highperformance image processing system
for image segmentation, detection &
classification of endometrium cancer is
very important
Materials and Tools
Matlab 2014
Database: mini-MIAS
Chapter
2
THEORETICAL BACKGROUND AND RELTED WORKS
This chapter aims to give a description of the
theories leading to the detection of uterine
abnormality
Gynecologic cancers
Gynecologic cancers begin in different
places within a woman’s pelvis, which is
the area below the stomach and in
between the hip bones.
Anatomy of the
Endometrial
Diagrammatic representation of the female anatomy, showing the uterine cavity, cervix
and vagina and the position of the tubes and ovaries
Endometrial cancer diagnosis
Pelvic examination
01
Neptune is the farthest planet
from the Sun and a gas giant
Pap smear
(may detect cancer spread to
cervix)
02
03
Jupiter is a gas giant
and also the biggest planet of
them all
Endometrial sampling
(hysteroscopy) or curettage is
mandatory
Transvaginal
ultrasound
04
Endometrial cancer



It is the most common gynecological cancer
It occurs most often in postmenopausal women, with less than 5%
diagnosed under 40 years of age
There is no effective screening program, but occasionally cervical
smears contain endometrial cells or double ultrasound endometrial
thickness of 4 mm or more indicating the need for endometrial
sampling
Epidemiology
• Endometrial cancer is the most common gynecological
malignancy in the West, but in India, the incidence rates are low.
• Most of the cancers present at an early stage and are associated
with a good prognosis.
Epidemiology
• Median age at diagnosis = 61 years
• 20% before menopause
• 5% before 40 years of age
Endometrial Cancer – Disease Burden
New Endometrial Cancer Cases
Libya ~1,32,000
World ~ 4,93,000
ILibya ~27%
Deaths due to Endometrial cancer
Libya ~ 74,000
World ~ 2,73,000
India -~27%
27%
Libya
Rest of World - 73%
Rest of World - 73%
Libya ~27% of new
Cervical Cancer cases in world
Rest of World - 73%
Libyaa ~27% of deaths
due to Cervical Cancer in world
20
Risk factors for
endometrial cancer






Age
Family history of endocrinerelated cancers (breast,
ovary)
Previous breast or ovarian
cancer
Endometrial hyperplasia in
the past
Radiation therapy to the
pelvis
High number of menstrual
cycles (early menarche,









Nulliparity
Infertility or failure of
ovulation
Unopposed estrogen
therapy
Tamoxifen treatment
Diabetes
Obesity
Sedentarism
Metabolic syndrome
Diet high in animal fat
Transvaginal Ultrasound - Purpose To Perform
The image of the internal organs is produced through ultrasound tests. Imaging tests like these can
help in finding out any abnormality in the organ, so, you can get the right treatment on time.
Transvaginal Ultrasound is different from the normal ultrasound as it is an internal examination
instead of the external one. Transvaginal simply means through the vagina and is done using a probe
that gets inserted in the vaginal canal to about 2 to 3 inches deep.
Technique
• Transvaginal sonography gives a more
detailed evaluation of pelvic architecture
using higher-frequency transducers at closer
proximity to pelvic structures.
Transvaginal Sonography
Types of sonograpgy
anterior
anterior
left
right
cephalad
cephalad
posterior
Important findings of Transvaginal ultrasonography
N=Normal
EC=Endometrial Cance
P=Endometrial polyps
TVUS images.
Literature review
Noise removal
(Thakur et al.et al., 2005)
comparative study of different noise reduction methods based on wavelet filter according
to different threshold values applied to ultrasound images
(Thakur et al.et al., 2005)
conducted one of very important denoised techniques to improve diagnostic information
form ultrasound, the Wiener filtering technique was used to remove speckle noise from
ultrasonic images of the liver. The search results of the algorithm was very useful for
denoising
Literature review
Image segmentation
Rawat and Gupta (2018)
proposed a technique that combines fuzzy C means and Darwinian particle swarm
optimization (PSO). Among fuzzy-based clustering algorithms, the FCM algorithm is the
most popular
Saravanan and Sathiamoorthy (2018)
developed a computerized segmentation technique based on active contours without
outline techniques for an effective PCOS classification of 3D ultrasound image
Literature review
Endometrial cancer detection
(Mrs. Snehal R. Shinde et al., 2019)
explored system of Endometrial Cancer Diagnosis using CAD systems is useful to improve
the present methods of diagnosis to obtain an accuracy of 87.5%
(Xue Wang et al., 2022)
conducted anther very important study In this study, 85 cases of three-dimensional
transvaginal ultrasound (3D TVUS) images were collected retrospectively, including 75
cases of endometrial adhesion and 10 cases of non-adhesion
PROBLEM
SOLUTION
Mercury is the closest
planet to the Sun and the
smallest one
Jupiter is a gas giant and
the biggest planet in the
Solar System
Chapter
3
THEORETICAL BACKGROUND AND RELTED WORKS
The aim of this chapter is to undertake a review of
digital image processing. This chapter discussed Image
processing techniques used in our work, this will cover
the fundamentals of three image processing methods:
image filtering, image edge enhancement and image
segmentation
Ultrasound digital image
A digital image is a numerical representation of a two-dimensional image, i.e., it is a discrete fu
nction. The digital image is described by discrete points, called pixels. The pixels are arranged i
n a grid and each pixel has its position, represented by the space coordinates, and color. The col
or is also discretized and its values are natural numbers between 0 and 255. In the case of digita
l grayscale images, a pixel with the value 0 represents a black pixel, the pixel with the value 25
5 represents white
Noise in images
• Images often degraded by random noise
– image capture, transmission, processing
– dependent or independent of image content
• White noise - constant power spectrum
– intensity does not decrease with increasing frequency
– very crude approximation of image noise
• Gaussian noise
– good approximation of practical noise
• Gaussian curve = probability density of random variable
– 1D Gaussian noise - µ is the mean
–  is the standard deviation
Gaussian Noise
Gaussian Noise, also known as Gaussian Distri
bution, is a statistical noise with a probability d
ensity function equal to the normal distribution
Salt-and-pepper noise
This type of noise which called Salt-andpepper noise is a type of noise usually
seen on digital images. It is also known as
impulse
Speckle Noise
Is a granular "noise" that inherently exists
in and degrades the quality of the active
radar, synthetic aperture radar (SAR),
medical ultrasound and optical coherence
tomography images.
Restoration in the Presence of
Noise
Spatial Filtering
Restoration in the Presence of Noise Only - Spatial Filtering
Image restoration is used to carry out several useful
tasks in digital image processing. One of the very
important tasks is noise removal using image filtering
which is a technique for modifying or enhancing an
image, when a filter is used to reduce the amount of
unwanted noise in the ultrasound image. a filter usually
operates on a neighborhood of pixels in an image. In the
spatial domain, image filtering is done by convolving
the raw image with the filter function to obtain the
filtered image, where the convolution takes place over
the neighborhood of each input pixel A mean filter and
a median filter are both types of filters that can be used
for noise removal. Whereas the mean filter is good
example of a linear filter, the median filter is an
example of a nonlinear filter
Mean Filters
Ordered-Statistic Filters
Adaptive Filters
Mean Filters
Performance superior to the filters discussed in Section 3.6 Degradation Model:
g ( x, y )  f ( x, y )  h ( x, y )   ( x, y )
To remove this part
Arithmetic Mean Filter(Moving Average Filter):
Computes the average value of the corrupted image g(x,y)
The value of the restored image f
1
fˆ ( x, y ) 
g ( s, t )

mn ( s ,t )S xy
mn = size of moving window
(P. Harikanth)
Order-Statistic Filters: Revisit
Original image
Subimage
Statistic parameters
Mean, Median, Mode,
Min, Max, Etc.
Output image
Moving
Window
Order-Statistics Filters
Median filters:
Are particularly effective in the presence of both bipolar
and unipolar impulse noise
fˆ ( x, y )  median g ( s, t )
( s ,t )S xy
(P. Harikanth)
Median Filter: How it works
A median filter is good for removing impulse, isolated noise
Salt noise
Pepper noise
Median
Moving
Window
Degraded
image
Salt noise
Pepper noise
Sorted
Array
Filter output
Normally, impulse noise has high magnitude and is
solated. When we sort pixels in the moving window,
noise pixels are usually at the ends of the array.
(P. Harikanth)
Therefore, it’s rare that the noise pixel will be a median value.
Chapter
4
EXPERMENTION OF ENDOMETRIAL CANCER DETECTION
This chapter describes the implementation of the
segmentation process of endometrial region form
uterine ultrasound image They were image
acquisition, image prepressing, image segmentation,
feature extraction, and classification.
Chapter
5
EXPERIMENTAL RESULTS
Chapter
6
FURTHER RESEARCH SCOPE
There is always more to work on…
in research
Thank you for your time and attention!
?
Questions?
(Comments)
Download