Multimedia signal processing and machine learning

Supported by the members of the Centre for Multimedia Computing, Communications, and Artificial Intelligence Research (MCCAIR), we aim to conduct world-class theoretical and applied research in the domains of;

  • multimedia signal processing;
  • machine learning; and
  • distributed computing.

We will be developing machine intelligence for the advanced and automated management and analyses of the growing volume of crime scene investigation (CSI) and cybersecurity related data that are embedded in digital multimedia and communications. IFS research demands complementary expertise in four core domains of ICT—signal processing, computational intelligence, communications, and information theory. MCCAIR has strong research track record in all these four domains, which is unique in Australia.

MCCAIR is a national leader in IFS research, analysing challenging theoretical problems, developing cutting-edge techniques and technologies, providing research training opportunities, engaging with the law-enforcement agencies and industries, partnering with overseas research centres, contributing to the national science & innovation agenda, and ultimately, building a safer Australia.

Research projects

Application-focused video coding for compressed-domain video analysis

Information-rich video contents are now effectively used for a large number of “non-viewing” applications in computer vision, video telemetry, photogrammetry, information forensics, and cybersecurity. The massive growth in video data repository has also created demand for meta-applications of video data such as automated video indexing and cataloging for browsing and query-based retrieval. The Project aims at developing a novel video coding paradigm, which first partitions video data into optimal cuboid-shaped segments, based on user-defined objective functions such as polarising colours and/or textures, and then encodes them independently with additional meta-data on features defined in the objective functions. The outcome of the Project will enhance Australia’s cybersecurity capability and directly benefit law-enforcement agencies, traffic management, remote health services, and production line monitoring.

Mass estimation for anomaly detection

We reveal that a phenomenon in machine learning algorithms which is in contrary to the conventional wisdom---more data the better. A 2016 Machine Learning Journal paper provides the theory and empirical evidence that good performing (distance-based) anomaly detectors can be trained using small data sets, and data more than the optimal data size will produce sub-optimal anomaly detectors.

Reinforcement learning for smart object tracking in CCTV

Law enforcement agencies are increasingly reliant on networks of surveillance devices such as CCTV to rapidly identify crimes as they are occurring, or as evidence for identifying, locating, and prosecuting criminals. This project will develop intelligent software agents based on reinforcement learning to control networks of smart cameras to enable the tracking of one or multiple target individuals.

Monitoring Internet-of-Things (IoT) objects in smart environment

Internet-of-Things (IoT) is increasing being used in every aspect of human life starting from advancing the global economy to improving the quality of life, productivity, and security. Monitoring IoT objects, which can be considered both beings and things, along with tracing their behaviour and activities in the cyber as well as physical worlds is a major research challenge. This project aims to investigate the development of a framework for monitoring and tracing the abnormal behaviour of IoT objects using the footprints left in both cyber and physical worlds. The context of the environment in which an IoT object resides will be considered as one of the parameters in the framework. The outcome of the project will be useful in diverse applications including smart cities and infrastructure, healthcare, surveillance and security.

Further information

Centre for Multimedia Computing, Communications, and Artificial Intelligence Research (MCCAIR)
Professor Manzur Murshed

Phone: +61 3 5122 6467
Email: manzur.murshed@federation.edu.au