Research Seminar - Adam Gatt & Sebastien Wong

[Elyse Perin]

Dear Colleagues

 

You are invited to attend a research seminar by Adam Gatt & Sebastien Wong

 

Date: Friday 9^th November 2012

 

Time: 12.00pm

 

Location: D1-05, Mawson Lakes

 

~ A light lunch will be provided ~

 

Adam Gatt,  PhD Candidate

Sys Arch

Supervised by Associate Professor David Kearney

 

 

Title:

Multi-object visual tracking without a prior knowledge

 

Abstract:

This paper presents CACTuS-FL, an innovative visual tracking system that combines feature- and shape-learning capabilities with multi-object tracking. CACTuS-FL is capable of probabilistically learning the feature responses, shape and motion parameters for multiple objects of interest simultaneously. The system converges towards regions of temporally consistent and spatially correlated local saliency, allowing for the automatic detection and tracking of interesting objects with minimal a priori knowledge required from the user. This capability is demonstrated against a range of video sequences, resulting in successful tracks for several diverse objects of interest. A comparative evaluation is presented to show that this generalizability comes at little to no cost to tracking accuracy or robustness.

 

Adam Gatt is a PhD student in the Reconfigurable Computing Laboratory at the University of South Australia. His research interests involve visual tracking, feature learning and tracker evaluation. Adam received a bachelor's degree in Information Technology with first class honours in 2008, and was awarded the Joyner Scholarship and UniSA Vice Chancellor and President's Scholarship in 2009.

 

Sebastien Wong joined DSTO from industry in 1999 and he currently leads Australia’s research effort into algorithms and architectures for electro optic threat warning systems. In 2008 Sebastien was awarded the Chief of Air force Gold Commendation for his work in the improvement of the AN/AAR-60 missile warning algorithms. Sebastien has produced 20 peer reviewed publications. Sebastien holds a bachelor’s degree in Computer Systems Engineering (with honours) from Curtin University, as well as a master’s degree in Electronic Systems Engineering and a PhD in Computer Science, both from the University of South Australia, where he is an adjunct Senior Lecturer. His research interests include algorithms and parallel processing architectures for autonomous visual tracking systems.

 

 

Kind regards

Elyse

 

 

 

 

 

 

 

 

[Elyse Perin]