Qualitative spatial reasoning for activity detection in video: Detecting relevant objects and their interactions

Hajar Sadeghi-Sokeh (ANU)

ARTIFICIAL INTELLIGENCE SEMINAR PhD Monitoring

DATE: 2013-10-30
TIME: 12:00:00 - 12:30:00
LOCATION: RSISE Seminar Room, ground floor, building 115, cnr. North and Daley Roads, ANU
CONTACT: JavaScript must be enabled to display this email address.

ABSTRACT:
Describing actions containing humans interacting with different objects and other humans is an interesting field of research. It involves recognizing a particular action that occurs in the video. In this regard, a good representation for modeling the behavior of interacting objects over space and time is essential. We focus on recognizing these activities using a qualitative representation that captures spatio-temporal aspects such as topology, size, distance and direction. By considering these as feature descriptors machine learning algorithms can be used to cluster or classify activities. Since determining the relevant object is very important in this process, one of our other main contributions is selecting the most relevant object to the activity from all possible moving objects from each video.
BIO:
PhD Start Date: 29/4/2011 (Expected) End Date: 28/10/2014 Supervisor(s): Jochen Renz, Stephen Gould Panel Members: Jochen Renz (Chair), Stephen Gould (Co-Supervisor), Lars Peterson (Advisor)