Hamish Pratt

Hamish Pratt

Higher Degree by Research Candidate

HDR Student

School of Computer Science

Faculty of Sciences, Engineering and Technology

I am a HDR student with the Australian Institute for Machine Learning and Visual Physiology and Neurobotics Laboratory at the University of Adelaide. I am currently researching motion detection in neural networks and how bio-inspired approaches for motion detection can help with object detection.

I am interested in the comparisons between biological systems and machine learning. The new wave of machine learning has opened up this 'new brain' that we can teach, but I want to know what this new brain can teach us.

PhD Project:
Neural networks such as YOLO have achieved state-of-the-art results for object detection by processing static frames. These networks are able to process videos by taking in frames continuously, however the larger number of frames requires more computational power. Insects such as dragonflies are expert hunters and use motion cues to identify targets with inconceivably small brains. Inspired by these computationally cheap biological systems, my research aims to combine motion with conventional object detection to make quick and accurate object detection algorithms.

  • Conference Papers

    Year Citation
    2020 Pratt, H., Evans, B., Rowntree, T., Reid, I., & Wiederman, S. (2020). Recurrent Motion Neural Network for Low Resolution Drone Detection. In 2020 Digital Image Computing: Techniques and Applications, DICTA 2020 (pp. 1-7). online: IEEE.
  • Position: HDR Student
  • Email: hamish.pratt@adelaide.edu.au
  • Campus: North Terrace
  • Building: Australian Institute for Machine Learning, floor 2
  • Org Unit: Australian Institute for Machine Learning - Operations

Connect With Me
External Profiles