Anomaly Detection in Lung CT Scans through AWS SageMaker: A Machine Learning Approach
Utilizing the National Lung Screening Trial (NLST) CT scans dataset as a primary resource, the project will use AWS SageMaker's scalable machine learning capabilities to develop, train, and deploy models that can perform detailed analyses of CT images. These models will be designed to learn from the data, improving their accuracy over time through continuous training and validation against known outcomes. The project's approach combines the scalability and power of cloud computing with the precision of machine learning.
-Develop an Anomaly Detection Framework: Utilize AWS SageMaker to create a scalable machine learning framework capable of analyzing cell and lung CT scans to detect anomalies, trends, and patterns indicative of lung diseases.
-Accuracy and Validation Enhancement: Implement machine learning algorithms and techniques to improve the accuracy of anomaly detection in lung CT scans, with continuous validation against a subset of the NLST dataset to ensure reliability.
-Interpretability and Explainability: Ensure the machine learning models are not just black boxes but provide interpretable and explainable results that can be understood and trusted by medical professionals.
-Integration and Deployment: Seamlessly integrate the developed ML models into existing healthcare workflows, allowing for easy access and use by radiologists and healthcare providers for second opinions or preliminary screenings.
Vincent Do vincentdo12@berkeley.edu
CloudatCalifornia is a club at UC Berkeley which is doing this research to build the project along