Logo equipe IVC

REVAMP, a franco-singaporean project

Visual attention modelling and applications are important research topics in media processing. Both the Singapore and France teams have substantial interests in the said areas and will work together toward more impactful research work, based upon the complementary technical skills available on the two sides.

With the fundamentals and rapport built during this project, we aim at major joint funding applications in Singapore, France and Europe in the near future. In particular, the project work targets several new applications which are expected to be deployed on large commercial scale in foreseeable future with several companies working in related fields.

Thus, this project is of high relevance to industries. In the light of this, we expect a significant interest and as a result major funding from companies which are active in the said areas.

REVAMP PROJECT

With the advancement of technology, large amount of visual information is produced in the form of images, videos, graphics and so on. However, current visual information processing systems suffer from several drawbacks.

The most important of these is the disregard of the way the human visual system (HVS) functions. As a consequence, processing is more mechanical (no discrimination between important visual signal information) and quite complex (visual data is processed in its entirety). To tackle these major shortcomings, we propose the use of visual attention (VA) for more efficient visual signal processing. In particular, we target several upcoming and challenging applications of VA in :

  • High dynamic range (HDR) content processing,
  • Image retargeting,
  • 3D media processing,
  • Graphics processing and so on.

Given the diverse nature and requirements of these applications, our proposed solutions will be based upon the latest findings in physiology, psychology, vision research, and neuroscience, as well as our existing substantial work in VA based perception modeling.

Carefully planned subjective tests will also be conducted to gain insights into intelligent incorporation of VA in the mentioned applications. Our joint project work is both novel and significant since the applications targeted by us are potential candidates of large scale practical deployment.

Hence our work is of immediate relevance to industries and academia alike. In essence, the REVAMP project will enable processing of important visual information more precisely resulting in better utilization of resources as well as providing better visual experience to the end user.

There is no report of similar work, and therefore the proposed joint project is expected to create new knowledge in this multi-disciplinary field. Due to the fundamentality and importance of the topic for visual information processing tasks, significant impact in terms of both theory and implementation is anticipated.

 

 

Logo-cnrsLogo-AtlanSticLogo-UniversiteNantesLogo-MinesNantesLogo-CentraleNanteslunamCOMUE