Visual optimization and compression to maximize the efficient storage of media.

A method for achieving the highly efficient compression of images in a lossy manner while maintaining a pre-set standard of visual quality.

 

The Technology

Visual quality of online images is rapidly increasing in importance with databases such as Facebook and Instagram becoming individual’s sole photo albums. To maximise the storage efficiency compression techniques must balance demand for quality and space.

UNSW has developed a method for lossy compression of images which reduces the visual distortions at a given bit-rate or reduces the bit-rate for a given level of distortion. This technique provides an optimised level of compression while maintaining a set standard of visual quality thereby providing the optimum compression or optimum bit-rate as required while minimising negative visual effects. The method may also be used for collections of images ensuring more sensitive images receive a larger allocation of the total data budget than those less sensitive.

 

Key Benefits

  • Ability to set an optimum bit-rate or visual quality level as the standard for optimum compression
  • Advantages over existing techniques with huge potential for efficiency increases in the social media market.
  • Measures visual distortion with respect to the human visual system
  • Either better quality for a set bit rate or best bit rate for a set quality

 

Applications

  • Efficient online image hosting minimising visual effects due to high levels of
  • compression (e.g. social media applications)
  • Preserving a standard of visual quality in online image databases (e.g. historical collections) while achieving maximum compression. 

  

The Opportunity

NewSouth Innovations is currently looking for partners to licence the technologies developed by Prof David Taubman, the inventor of the successful Kakadu Software JPEG2000 compression SDK. Due to the similarities across the portfolio these technologies are available as a group or individually for use in potential licensee’s commercial products.

Click here to download a PDF version of this summary and additional information.

 

Patent Information:
For Information, Contact:
Daniel Gronowski
Business Development Manager
University of New South Wales
d.gronowski@nsinnovations.com.au
Reference ID: 99_0114
Inventors:
David Taubman
Keywords:
Software
© 2025. All Rights Reserved. Powered by Inteum