LUNGS CANCER DETECTION USING DEEP LEARNING

Lung cancer is the major and biggest evolving disease and is the top reason for cancer-related deaths within the US and in the world. It has an increasing rate of affected patients and evolving patient checkups on daily basis all around the world. Our product is a web-based cancer detection s

Project Title

LUNGS CANCER DETECTION USING DEEP LEARNING

Project Area of Specialization

Artificial Intelligence

Project Summary

Lung cancer is the major and biggest evolving disease and is the top reason for cancer-related deaths within the US and in the world. It has an increasing rate of affected patients and evolving patient checkups on daily basis all around the world.

Our product is a web-based cancer detection software algorithm for people infected by lung cancer.

This algorithm works as an alternative to traditional cancer detection methods which include using medical x-rays images and concerning lungs specialists to evaluate the detection results. The model involved is trained by using thousands of normal and pneumonia x-rays images to predict and evaluate the given user x-ray image with the highest accuracy possible all in a matter of a few clicks.

Project Objectives

Our main objective is to help people detect lung cancer remotely free of cost and within the comfortable environment of their home in a matter of a few clicks on our website.

People can immediately get desired results.

Patients can evaluate the cancer stage rapidly.

We have deployed a highly intelligent deep learning algorithm on the web that can help a vast amount of audiences over the internet diagnose their cancer by mere input of images.

Project Implementation Method

The proposed automated lung nodule identification and categorization system employs a variety of methods. Physiological symptoms, CT scan analysis, and clinical biomarkers are used to make judgments by the system. With a family history of lung cancer, unattended physiological symptoms can lead to a lung cancer prediction. Although deep learning-based CT scan analysis approaches beat radiologists in detecting lung nodules, particularly those with a diameter of less than 6 mm, distinguishing between benign and malignant nodules is a substantial and difficult issue due to the massive overlap of features. The suggested system makes judgments depending on different approaches to minimize the negative predictive value. For the categorization of early-stage lung nodules, clinical indicators, particularly plasma proteins and blood tests, are extremely relevant.

Benefits of the Project

It is advantageous to be detected with lung cancer at an early stage since therapy may then be started to prevent the disease from becoming detrimental. As a result, this work offers a comprehensive review of several machine learning algorithms for classifying lung cancers using CT scans or X-ray pictures. Many classifiers have been utilized by researchers in the literature. All of the research that used Deep Learning approaches yielded excellent accuracy results, with (Li et al., 2020) employing multi-resolution patch-based CNN's achieving the highest result of around 99 percent.

Technical Details of Final Deliverable

• The model developed will be able to respond to any input in clear format (jpg or png) and visuals. If the format is wrong or the image is broken somehow, the website should return an error saying, “please try another image”.
• If the model understands the input, it will fetch that input and perform techniques of deep learning algorithm embedded in the backend of the website.
• The model will be able to analyze the input image and provide a “cancer score” accordingly with maximum and utmost accuracy.
• The evaluated result will be displayed on the website to the end-user along with an accuracy score and a statement explaining the seriousness of the matter like a cancer report.
• The backend program will be able to always function properly. Its working should not be affected by any kind of bug.
• The backend support will be able to read the image inputs given by the user and process them in a sufficient amount of time by reducing time complexity.
• If there comes an error at the time of replying, the backend should force the algorithm to generate a default message such as “sorry, I didn’t understand” or “please come again” instead of giving a wrong and non-relevant response.
• If any question is asked by the user, that is out of the scope of the model, at that time the model should simply say like “sorry, but I don’t know the answer” or “sorry, but you asked something different that I don’t know”.
• A response generated will be short, easy to understand, and to the point to boost client/user satisfaction.
• Every function available on WEBSITE should be properly linked with the backend.

Final Deliverable of the Project

Software System

Core Industry

IT

Other Industries

Medical , Health

Core Technology

Artificial Intelligence(AI)

Other Technologies

Sustainable Development Goals

Good Health and Well-Being for People

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com