Site-wide links

<<   February 2016   >>
Su Mo Tu We Th Fr Sa
w   1 2 3 4 5 6
w 7 8 9 10 11 12 13
w 14 15 16 17 18 19 20
w 21 22 23 24 25 26 27
w 28 29          

  Search By:
Date Range:  
Start Date Click Here to open the calendar
End Date   Click Here to open the calendar
include date range

Event Details
Event Name: GCCIS PhD Colloquium Series
Category: Academic/College Events
Sub Category: GCCIS
Description: Robust Learning-Based Collimation Detection in Digital Radiographs

Presentation by
Hongda Mao

X-ray radiography is frequently used in clinical examinations. In the exams, the radiographers often use collimation to the appropriate anatomy of interest to minimize the overall integral dose to the patient. The shadow regions of the image are not diagnostically meaningful and could impair the overall image quality. Thus, it is desirable to detect and exclude the shadow regions (collimation) before image display. In this presentation, after a brief overview of the state-of- the-art of X-ray imaging techniques, we will propose a robust learning based approach for collimation detection in digital radiographs. Specifically, a random forests learning based region detector is used to provide pixel-wise image classification and each pixel is labeled as either in-collimation or out-of-collimation. At the same time, a discriminative, learning-based landmark detector is utilized to detect and localize the corners of the collimation in the image. The final detection result is obtained by combining the results from region and landmark detectors. We evaluate our algorithm in a database with 665 X-ray images in a wide variety of types and dosages, and obtain better results compared to the existing methods.

Hongda Mao is a sixth year Ph.D. student in the Ph.D. Program in Computing and Information Sciences at RIT. Before coming to the US, he received his B.S. degree in information engineering in 2006 and M.S. degree in optical engineering in 2008 at Zhejiang University, Hangzhou, China. He worked as a research intern at Carestream Health (formerly Kodak's Health Group) and Siemens Healthcare in 2012 and 2013 respectively. He is currently working on his thesis under the supervision of Dr. Pengcheng Shi. His research interests include computational cardiology, computer-aided diagnosis, robust image segmentation and machine learning.
Schedules: 11/01/2013   (11:00 AM - 12:00 PM)

Contact: Joyce Hart
Phone: 475-6193
Cost: Cost - Free