Speaker: 

Ruchi Guo

Institution: 

UC Irvine

Time: 

Monday, October 5, 2020 - 4:00pm

Host: 

Location: 

Zoom

Electrical impedance tomography (EIT) is a very promising technique for non-invasive, radiation-free type of medical imaging. But a high-quality reconstruction for the EIT problem is challenging due to its severe ill-posedness. In this talk, we present two frameworks to construct deep neural networks to solve EIT problems based on the idea of direct sampling methods (DSMs) proposed by Chow, Ito and Zou in a series of works. Specifically we propose the fully-connected-neural-network and convolutional-neural-network deep direct sampling methods (FNN-DDSMs and CNN-DDSMs). Both the DDSMs are able to efficiently reconstruct buried inclusions with high accuracy and with only very limited boundary Cauchy data pairs. As a remarkable feature, the reconstruction is highly robust with respect to large noise (up to even 20%). 

Topic: MATH 298A SEM A: APPLIED MATHEMATICS (45300)
Time: Oct 5, 2020 04:00 PM Pacific Time (US and Canada)

Join Zoom Meeting
https://uci.zoom.us/j/91243957681?pwd=L2svcDNJOGowclk5WFAzenJDUHh6UT09

Meeting ID: 912 4395 7681
One tap mobile
+16465588656,,91243957681# US (New York)
+13017158592,,91243957681# US (Germantown)

Dial by your location
+1 646 558 8656 US (New York)
+1 301 715 8592 US (Germantown)
+1 312 626 6799 US (Chicago)
+1 669 900 6833 US (San Jose)
+1 253 215 8782 US (Tacoma)
+1 346 248 7799 US (Houston)
Meeting ID: 912 4395 7681
Find your local number: https://uci.zoom.us/u/acihMZWzg

Join by Skype for Business
https://uci.zoom.us/skype/91243957681