"Low-Complexity Driving Event Detection using Information Extracted from the Video Encoding Process"

Dr. Enrico Masala, Assistant Professor, Computer Engineering, Politecnico di Torino, Italy

October 24th (Thursday), 2:30pm
Harold Frank Hall (HFH), Room 4164

This talk discusses how to use a smartphone equipped with a camera to detect driving events in a vehicular environment. A standard video compression algorithm is applied to the video captured by a camera facing the road, and a set of interesting and easy-to-extract features is identified and extracted from the video encoding process with low complexity. A machine learning algorithm based on SVM, after appropriate training over several hours of video annotated by a human operator, shows that it is possible to identify several events with a very good identification rate. A potential application of such a system is, for instance, the possibility to equip any type of vehicle, regardless of the age and model, with a simple device, e.g., a smartphone, that can warn the driver of potentially dangerous situations in real-time.

About Dr. Enrico Masala:

photo of enrico masala Enrico Masala received the Ph.D. degree in computer engineering from the Politecnico di Torino, Turin, Italy, in 2004. In 2003, he was a visiting researcher at the Signal Compression Laboratory, University of California, Santa Barbara, where he worked on joint source channel coding algorithms for video transmission under the supervision of Prof. Rose. Since 2010 he is Assistant Professor in computer engineering at the Politecnico di Torino, Italy. His main research interests include algorithms for multimedia processing (especially video), simulation and performance optimization of multimedia communications over wireline and wireless packet networks.

Hosted by: Professor Kenneth Rose, Signal Compression Lab