Statistics Seminar - 10/18/22

Abstract: 

Early-stage breast cancers are very challenging. One reason may be the lack of temporal analysis. However, a radiologist usually uses the current and prior mammograms side by side to evaluate changes over time. We propose a computer-aided detection (CAD) method for breast cancer screening using recurrent neural network (RNN), convolutional neural network (CNN), and follow-up scans. First, mammographic images are examined by three cascading object detectors to generate a region proposal. Then all regional images fed to a pre-trained CNN (VGG-19 model) to extract features. The image features come from registered prior scan, current scan, and their difference image. Finally the image features are combined to train a RNN classifier. The RNN functions as a temporal analysis, which can factor in multiple follow-up scans. The experimental results show that the performance of the proposed CAD method is very promising.

Biographical Information: 

Dr. Yufeng Zheng has joined the University of Mississippi Medical Center faculty as an associate professor of Data Science in the John D. Bower School of Population Health since January 2021. He currently serves as a Graduate Program Director in the department.

Recipient of the B.S. in optical engineering in 1989, the M.S. in optical engineering/image processing in 1994 and the Ph.D. in optical engineering/image processing in 1997 from Tianjin University (one of 39 Project 985 Universities in China), Zheng served as a senior software engineer at Huawei Tech Company in Beijing from 1997-99 and senior CT imaging algorithm engineer at GE Hangwei Medical Systems Company in Beijing from 1999-2001 before moving to the U.S. A postdoctoral research associate at the University of Louisville, Kentucky, from 2001-05, Zheng joined the Alcorn State University faculty as an assistant professor in 2006 and earned promotion to associate professor in 2012 and professor in 2019.

Patent holder of a face recognition system and method using face pattern words and face pattern bytes, Zheng has been well funded ($2.4M total) as the PIs of several research grants. The author or coauthor of three books, six book chapters, 25 articles in peer-reviewed journals and 60 journal papers, Zheng’s research interests include image procession and pattern recognition; neural network and artificial intelligence; information fusion (image/score fusion), biometrics (facial recognition); machine learning and computer vision; and computer-aided diagnosis (medical imaging).