Center for Ocean-Atmospheric Prediction Studies,
Florida State University
"Applied deep learning and scientific visualization in geosciences and medical imaging"
Nov 30, 2021 Schedule:
- 2:30 to 3:30 PM Eastern Time (US and Canada)
Meeting # 96985725848
Abstract:
This talk will summarize some projects that I have worked on over the last five to seven years. In geosciences, specifically in oceanography, I will discuss the problem of data assimilation in operational ocean models and the challenges of using deep learning methods to speed up this process. Likewise, I will examine the machine learning methods used to predict the ocean’s interior information from surface data. Regarding atmospheric sciences, I will discuss an operational data-driven air quality forecast for Mexico City. Then, I will move on to the field of medical imaging, where I will discuss our contributions in the field of automatic organ and tissue segmentation, specifically the work done for prostate zones and prostate cancer, kidneys, and organs at risk of cervical cancer. I will illustrate the significant advances in this field thanks to deep learning and demonstrate how sensitive these methods are to changes in the training data.