Methods of classification to boost breast cancer screening efficiency include:-
Methods of classification to boost breast cancer screening efficiency include:-
Abstract:-
Breast cancer has an extremely poor median survival rate and
is a very aggressive type of cancer. Due to the growth of malignant cells in
the breast, the number of deaths among women between the ages of 15 and 55 is
rising nowadays. It is the primary factor in female fatalities. Therefore, the
only way to improve is by early patient diagnosis. By enhancing the tools for
early diagnosis and analysis of breast cancer, machine learning (ML) approaches
can help doctors improve patient survival rates. Radiologists currently utilise
mammography as the most effective imaging technique for detecting breast
tumours.
Introduction:-
Today, breast cancer is one of the most prevalent cancers in
both men and women. Breast cancer is described by the malignant cells that develop
in the breast tissue. Breast cancer risks rise in correlation with factors like
female sex, obesity, lack of exercise, postponing childbearing, etc. Breast
cancer is brought on by the breast cells' atypical proliferation. These cells
grow quickly in both benign and malignant cancers. In benign tumours, cell
growth ceases at a specific stage, while in malignant tumours, cell growth
continues until the entire body is impacted.
Women frequently die from breast cancer, and the number of
people who have been diagnosed with the disease is rising. The likelihood of
recovery can therefore be increased by early diagnosis. The goal of the article
was to develop a novel efficient method for preprocessing that would allow for
feature extraction without eliminating pectoral muscles. In order to transform
the approximation matrix into 1-Dimensional using zigzag scanning and remove
volatile signal components, Farzam et al. first performed discrete wavelet
transform to the image. This paper's limitations stem from the sheer volume of
examples it requires.
Methods of
classification to boost breast cancer screening efficiency include:-
Breast cancer diagnosis with imaging has been widely used in
medicine. It's crucial to find breast cancers in their early stages. The suggested
early detection method is mammography. To diagnose the mammography breast
tumour, multiple radiologists are available in the diagnostic facilities. The
radiologists performed single readings with or without the aid of CAD
technologies as well as double readings in known and unknown ways. A second
reading is suggested for confirmation in order to ensure accurate results.
Double reading is a practise quality in the Dutch language.
The radiologists can categorise the outcomes into several categories and stages
after performing a second reading. It may produce erroneous positive results.
Because of their intricate structure, mammogram pictures are challenging for
radiologists to correctly categorise and extract characteristics from. Numerous
methods for feature extraction and disease classification have been developed
by several researchers, however there is still room for improvement.
It is possible to use several deep learning models to carry
out various tasks, including object detection, visual tracking, semantic
segmentation, and classification. To carry out the classifications, the
researchers suggested a variety of models, including AlexNet, GoogleNet,
ResNet, MobileNet, and EfficientNet.
Three
datasets were used to create this paper. Images are improved through data
augmentation. For the best feature selection, the Modified Entropy controlled
Whale Optimization Algorithm (MEWOA) is suggested. These are the main
contributions:-
1.
Three mathematical formulas—horizontal shift,
vertical shift, and rotation 90—are used to achieve data augmentation.
2.
Deep features are collected from the middle
layer (average pool) rather than the FC layer in two deep learning pre-trained
models, Nasnet mobile and MobilenetV2.
3.
For the purpose of selecting the best features
and minimising the computational cost, we suggested a Modified
Entropy-controlled Whale Optimization Algorithm.
The best deep learning features were combined using a
serial-based threshold method.
According to a WHO poll, 2.3 million women will be diagnosed
with breast cancer in 2020, and 685,000 people will die from the disease
globally. At the time of detection, 81 percent of women with cancer are over
50. Breast cancer is the second most common cancer in the world and the first
in India. India's survival rate of 66% is extremely poor when compared to the
90 percent in the United States and the 90.2% in Australia. However, there is a
90% or higher chance that this cancer will respond to treatment. To reduce the
fatality rate, it is therefore necessary to find cancer extremely early on.
There are numerous methods for detecting benign and
malignant tumours before symptoms show up in the healthcare industry, including
mammography, sonography, ultrasound, and MRI. Other continuing research
projects include PET (positron emission tomography) scans, thermography,
ductograms (ducto lavage, ductoscopy), etc. These studies are utilised to
classify breast cancer anomalies and serve as a second opinion for doctors.
DL-CADs (Deep learning CADs) are now in use and are superior to standard CADs
for complex data analysis. This article covered a thorough analysis of the deep
learning methods and data sets used to categorise breast cancer. resulting in
difficulties, restrictions, or unfinished business in this field of study.
Contact
Us:-
Reach out to us: https://breastcancerpathology.universeconferences.com/
Mail: pathology@universeconferences.com| info@utilitarianconferences.com |
breastcancer@ucgconferences.com
Whatsapp: +442033222718 Call: +12073070027
Previous
Blog Post Links:-
·
https://medium.com/@elizaedwards2021/breast-cancer-disease-f0324f19b8a2
Comments
Post a Comment