Location

https://ccse.kennesaw.edu/computing-showcase/cday-programs/spring2021program.php

Streaming Media

Document Type

Event

Start Date

26-4-2021 5:00 PM

Description

In the United states, 13% of women are diagnosed with breast cancer in their lifetime, and it is the second leading cause of death by cancer in women. Early detection and screening can result in an increase of life expectancy by 10 years on average. Unfortunately, breast cancer can be challenging to detect, since it can appear anywhere in the breast. Cancer that is detected in its early stages can give patients more options and save thousands of dollars in medical costs. Some of the most recent developments in computer science and machine learning are in the biomedical field, especially individualized healthcare. There is also an increase in the demand for telehealth options, reducing healthcare costs. With the help of computational technology, medical practitioners will be able to process data more quickly, which will allow more patients to have access to reliable treatment. Besides, systematic processes for interpreting various data types (such as clinical features, genetic information, and medical images) can identify trends that a human eye would not detect. This project aims to design and implement an artificial intelligence-based model called BreastNet to classify breast cancer into high and low-risk categories based on a combination of MRI images and clinical data. BreastNet uses a convolutional neural network (CNN), a type of machine learning methodology that imitates how the human brain learns information. Neurons fire in a connected pathway, reinforcing the relationship between a stimulus and the correct outcome. In this case, the CNN identifies characteristic features within the MRI that correspond to different life expectancy outcomes, which are notated in the clinical data. The clinical data serves as a loss function, which allows the network to identify how well the current model performs on images. We will evaluate the model by dividing the dataset into three partitions: training, validation, and testing, and then uses the evaluation metrics of Accuracy, Loss, F1 Score, Precision, Recall, Specificity, and Sensitivity.Advisors(s): Dr. Mohammed AledhariTopic(s): Artificial IntelligenceCS 4267

Share

COinS
 
Apr 26th, 5:00 PM

UR-46 BreastNet;

https://ccse.kennesaw.edu/computing-showcase/cday-programs/spring2021program.php

In the United states, 13% of women are diagnosed with breast cancer in their lifetime, and it is the second leading cause of death by cancer in women. Early detection and screening can result in an increase of life expectancy by 10 years on average. Unfortunately, breast cancer can be challenging to detect, since it can appear anywhere in the breast. Cancer that is detected in its early stages can give patients more options and save thousands of dollars in medical costs. Some of the most recent developments in computer science and machine learning are in the biomedical field, especially individualized healthcare. There is also an increase in the demand for telehealth options, reducing healthcare costs. With the help of computational technology, medical practitioners will be able to process data more quickly, which will allow more patients to have access to reliable treatment. Besides, systematic processes for interpreting various data types (such as clinical features, genetic information, and medical images) can identify trends that a human eye would not detect. This project aims to design and implement an artificial intelligence-based model called BreastNet to classify breast cancer into high and low-risk categories based on a combination of MRI images and clinical data. BreastNet uses a convolutional neural network (CNN), a type of machine learning methodology that imitates how the human brain learns information. Neurons fire in a connected pathway, reinforcing the relationship between a stimulus and the correct outcome. In this case, the CNN identifies characteristic features within the MRI that correspond to different life expectancy outcomes, which are notated in the clinical data. The clinical data serves as a loss function, which allows the network to identify how well the current model performs on images. We will evaluate the model by dividing the dataset into three partitions: training, validation, and testing, and then uses the evaluation metrics of Accuracy, Loss, F1 Score, Precision, Recall, Specificity, and Sensitivity.Advisors(s): Dr. Mohammed AledhariTopic(s): Artificial IntelligenceCS 4267