Biowulf 25th Anniversary Seminar Series: Dr. Sameer Antani (NLM)

About this Class

Please join us on Thursday for our third talk in the Biowulf 25th Anniversary Seminar Series:

Explorations in use of Synthetic Images for Biomedical ML/AI Research

This talk will be videocast at: https://videocast.nih.gov/watch=54838

Abstract:
The integration of synthetic images in biomedical machine learning (ML) and artificial intelligence (AI) research is revolutionizing the field by addressing critical challenges such as data scarcity, privacy concerns, and bias. This talk highlights our many research explorations into the development, use, and outcomes with synthetic data in ML/AI for biomedical research and applications. It will cover several of the latest advancements and applications of synthetic image generation techniques, including Generative Adversarial Networks (GANs) and Diffusion Models (DMs). It will address methodologies and challenges for creating high-fidelity synthetic medical images across various modalities, such as optical images, CTs and X-rays. The discussion will highlight how these synthetic images are utilized for data augmentation, model training, and validation, toward enhancing the robustness and generalizability of ML models – and resulting findings and open questions. Results from past and ongoing research will be discussed to provide the audience with a better understanding of the opportunities and pitfalls in this space.