Probabilistic / Deep Generative Models and Inference

To solve the inverse problem of electrophysiological imaging, we use the probabilistic graphical model and inference framework. We show that we can combine information from different sources by leveraging the power of graphical models and design good inference strategies. Also, we explore how we can combine the modern deep generative models with the traditional graphical models and inference strategies to solve the inverse problem.

References

  1. Ghimire, S., Sapp, J.L., Horáček, B.M. and Wang, L., 2019. Noninvasive Reconstruction of Transmural Transmembrane Potential With Simultaneous Estimation of Prior Model Error. IEEE Transactions on Medical Imaging, 38(11), pp.2582-2595.

  2. Ghimire, S., Dhamala, J., Gyawali, P.K., Sapp, J.L., Horacek, M. and Wang, L., 2018, September. Generative modeling and inverse imaging of cardiac transmembrane potential. In International Conference on Medical Image Computing and Computer-Assisted Intervention (pp. 508-516). Springer, Cham.

  3. Ghimire, S., Sapp, J.L., Horacek, M. and Wang, L., 2017, September. A variational approach to sparse model error estimation in cardiac electrophysiological imaging. In International Conference on Medical Image Computing and Computer-Assisted Intervention (pp. 745-753). Springer, Cham.

Sandesh Ghimire
Postdoctoral Researcher

My research interests include machine learning, computer vision and medical imaging.