Computer vision, robotics, concurrent and real-time systems
Brief Description of Objectives
Computer Vision Research at AI lab, UNSW
Vision is one of many sensory modalities which allow human beings to
perceive and understand the world around them. Computer vision, or
machine vision, attempts to extract, characterize and interpret information
from images of the real world, and thus duplicate human vision, using
computers.
This is a broad and interdisciplinary area. The focus of research in this
group is mainly middle and higher level vision, which deal with extracting
and labelling objects in images, and interpreting scenes. The group also
has interest in applying machine learning techniques to vision problems.
Projects
1. Generating symbolic descriptions of 2-D blocks world
2. Generating Symbolic descriptions of 3-D scenes
3. Learning object models from images
4. Learning motion models from image sequences
5. Human animation
6. Human facial expression recognition using neural networks
7. Vision for a remotely controlled robot arm
8. Obstacle avoidance using vision and ultrasonics
Robotics
Robotics is the study of the intelligent connection of perception to action.
Robotics research at this lab is focussed on algorithmic motion planning,
robot vision applications and intelligent robotics. Motion planning is
required when a robot needs to move from a source to a destination within
its environment. Vision is a frequently used sensing modality for a robot.
Under intelligent robotics, robots, both mobile and table-mounted, are used
to test hypotheses arising in our AI research, including machine learning
and modelling motion.
Robotics research in this department is focussed on
the higher faculties of vision and control and the application of
Artifical Intelligence and formal languages concepts to these areas.
Projects
1. Vision for a remotely controlled robot arm
2. remote control for a robot arm
3. Ultrasonics for a mobile robot
4. Obstacle avoidance using vision and ultrasonics
5. Prototype development for Autolab using Esterel
6. Motion planning for mobile robots
Control Applications
The use of formal languages in the analysis, design
and implementation of control systems is the focus of research in this area.
As is apparent, research and development go hand in hand in this area, and
a small group (of 1 staff member and a few students) are so engaged. Our
focus is on reactive systems at present, and real-time systems are also
of interest.
Projects
1. Control design of Autolab using the reactive paradigm
2. Logical Specification and Analysis of real-time scheduling systems
3. Statecharts simulation environment
Contact: Arcot Sowmya
Vision and Pattern Recognition is part of the Department of Artificial Intelligence, in the School of Computer Science and Engineering, Faculty of Engineering, University of New South Wales.