Amy Reibman - Video Quality in an Age of Universal Cameras
Start:
End:
Location:
138 DeBartoloHost:
Patrick Flynn
Phone: 574-631-8803
Website: http://www.nd.edu/~flynn
Office: 384A Fitzpatrick Hall
Curriculum Vitae
Affiliations
Digital cameras are everywhere, constantly recording our environment. Cameras
are distributed in hallways, classrooms, and on roadways. Wearable (body-mounted) cameras and vehicle-mounted “dash-cams” all record video without active intentional guidance by a human. Wherever the person goes or looks, video is captured. Unfortunately, the images captured from these cameras are not always high quality, even if they have a high spatio-temporal resolution. This reduces their effectiveness when they are subsequently watched by viewers or used by automated analytics algorithms to extract information.
In this talk, I’ll present two topics motivated by these new scenarios. On the first topic, I’ll consider perceptual quality of video recorded from First-Person (FP) (wearable) cameras, including a “mutual reference” quality estimation of images extracted from the video, and measurements of both the stability and First-Person Motion Information from the FP videos.
I’ll also demonstrate a stabilization algorithm that balances both stability and first-person motion. On the second topic, I’ll present a quality assessment tool for predicting the accuracy of a several commonly-used pedestrian detectors
in a practical system that includes video compression.
If time permits, I’ll discuss some ongoing research in video analytics for
real-world applications, where quality degradation is pervasive and labels
are difficult to obtain.
Seminar Speaker:
Amy Reibman
Purdue University
Amy R. Reibman is a Professor of Electrical and Computer Engineering
at Purdue University. She pursued 23 years of industrial research at AT&T Labs -- Research, where she was a Distinguished Member of Technical Staff and a Lead Inventive Scientist. Her research interests are in video analytics, video stabilization, and image and video quality assessment. She has also done pioneering work on video transmission over packet networks. She has published over 35 journal papers, 100 conference papers, and holds more than 40 US patents.
Dr. Reibman was elected IEEE Fellow in 2005, for her contributions to
video transport over networks. She was chair of the IEEE Fellow Committee
in 2016-2017, a Judge on the IEEE Fellow Committee from 2012-2015,
and a member of the IEEE Fellow Strategic Planning Committee from 2013-2018.
In 1998, she won the IEEE Communications Society Leonard G. Abraham Prize
Paper Award. She was the Technical co-chair of the IEEE International
Conference on Image Processing in 2002, and she is currently Technical co-chair
of the IEEE Global Conference on Signal and Information Processing in 2018.
She was a Distinguished Lecturer for the IEEE Signal Processing Society
from 2008-2009, was a member of the IEEE Awards Committee from 2010-2012.
