Every 60 seconds, more than 1, 5 millions Gigabytes of data are produced in the world, with a huge portion of images and videos. All the compression algorithms, including the most powerful ones, are overwhelmed by such a digital world explosion. Relying on the users’ accountability to limit this data volume burst might be limited as long as the only choice they have is to keep or delete their data.
In this project, we intend to offer to users a third choice called Data Repurposing. It consists in changing the initial data format for another type of visualization in a more compact representation, thus leading to drastic compression ratios. This enables the compression to withdraw tremendous amount of useless information while condensing the important information into a packed recycled signal. In a nutshell, in the data repurposing framework, the decoded signals target subjective exhaustiveness of the information, rather than fidelity as in the traditional compression algorithms.
This is a complete change of paradigm for image and video compression, which will enable gigantic compression gains. Data Repurposing lays the foundations for new generations of image and video compression algorithms. The aims of this project are i) to define this new compression scenario called Data Repurposing, suited for massive collections of image and video data ii) to evaluate the theoretical performance of such a paradigm and iii) to propose innovative coding approaches targeting unprecedented compression ratios for both image and video collections.
The project DARE rethinks the image and video compression and paves the road to the idea of using compression to limit the electrical energy consumption due to storage.