SHRP 2 NDS User Experience Webinar Series: Video Reduction & Analysis Methods- Part II, Automated Video Analysis
On Wednesday, September 13, 2017 from 12:00 PM to 1:00 PM EDT, the Virginia Tech Transportation Institute (VTTI) will host a webinar in a series that explores ways to use the second Strategic Highway Research Program (SHRP 2) Safety Data. The webinar “Video Reduction and Analysis Methods: Part II, Automated Video Analysis” will offer an overview of automated video reduction techniques. The SHRP 2 NDS User Experience Webinar Series is designed to provide a periodic opportunity for users to engage in discussions on topics suggested by the user community.
The Second Strategic Highway Research Program Naturalistic Driving Study (SHRP 2 NDS) was a three-year data-collection effort that produced continuous driving data for over 3,000 drivers, including thousands of safety critical events, including both crash and near crash events, as well as a large sample of baseline driving events.
Naturalistic driving studies such as the SHRP2 NDS collect large troves of data, including multiple streams of video data. Unlike sensor and questionnaire data often included in such studies, video data typically requires extensive research goal-specific manipulation to extract value and prepare the data for analysis. Recent computational advances have made automated approaches more feasible for some research needs. In this webinar, which compliments the previous webinar on manual coding methods, a suite of computer-vision and machine learning approaches to NDS analysis will be presented. The data analytic trade-off between volume of data, reduction, and the inherent noise and accuracy of the processing available will be explored. Specific tools developed for the FHWA Safety Training and Analysis Center (STAC) will be presented with discussions on their applicability as well as extensibility. Some test use cases will be presented. The complimentary nature of automated analysis and manual analysis will be explored as well to allow attendees to understand the possibilities of such methods.
The panel will include:
- Suzie Lee, Center for Data Reduction and Analysis Support, VTTI
- Tom Karnowski, Oak Ridge National Laboratory
The webinar host will be Suzie Lee, a Research Scientist in the Center for Data Reduction and Analysis Support at VTTI. Tom Karnowski of Oak Ridge National Laboratory will present. Tom managed the independent evaluation of a series of Federal Highway Administration Exploratory Advanced Research (EAR) projects, which focused on the use of machine learning and computer vision to perform automated feature extraction of several of the video streams available in the SHRP2 NDS.
The webinar host will collect specific questions throughout the webinar, and time will be allotted for questions and discussion after the panelists complete their presentations.
Date & Time: Wednesday, September 13, 2017 from 12:00 PM to 1:00 PM EDT
Instructions: To join the webinar on September 13, attendees can click or paste this link in a browser: https://goo.gl/vaeULV and follow the instructions. The webinar is open to anyone interested in this topic; no pre-registration is necessary. The link is also posted on the InSight Website.
A recording and accompanying slides for Part I of this webinar series, Video Reduction and Analysis Methods: Manual Reduction, are also available on the Background page of the InSight Website.
The SHRP2 NDS User Experience webinar series is designed to provide a periodic opportunity for InSight users to engage in discussions on topics suggested by the user community. If you have a topic that you’d like to hear about, let us know by emailing InSightFeedback@vtti.vt.edu.