Abstract:

In this talk, the speaker will present several biomedical image processing applications which can leverage the power of Deep Learning/Machine Learning to perform such tasks on image-to-image translation. For example, reconstruction, segmentation, and denoising. Additional tips and tricks in model and loss engineering will be also introduced. Furthermore, the data-centric augmentation to enrich the golden labeled datasets in the radiographical imaging field will also be briefly discussed.

About the Speaker:

TRAN MINH QUAN is currently an affiliated lecturer at VinUniversity, a member of VinGroup, the largest enterprise in Vietnam by capitalization. He got a B.S. degree in Electrical Engineering at KAIST (2012) and received a Ph.D. degree in Computer Science, focused on GPU Computing, at UNIST (2019), both from South Korea. His research interests are image-to-image translation, inverse problems, and high-performance visual computing, specializing in biomedical image processing and applications. Recently, he has been working on Deep learning-based Generative Adversarial Networks on various problems (neural volume rendering, variationally reconstructed radiographs, anatomical graphs). He also serves as a technical reviewer for Neurocomputing, IEEE Transactions, other journals, and conferences. He is a member of the AI Consultant Committee of Ho Chi Minh City, Vietnam.