English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In the current golden age of multimedia, human visualization is no longer thesingle main target, with the final consumer often being a machine whichperforms some processing or computer vision tasks. In both cases, deep learningplays a undamental role in extracting features from the multimediarepresentation data, usually producing a compressed representation referred toas latent representation. The increasing development and adoption of deeplearning-based solutions in a wide area of multimedia applications have openedan exciting new vision where a common compressed multimedia representation isused for both man and machine. The main benefits of this vision are two-fold:i) improved performance for the computer vision tasks, since the effects ofcoding artifacts are mitigated; and ii) reduced computational complexity, sinceprior decoding is not required. This paper proposes the first taxonomy fordesigning compressed domain computer vision solutions driven by thearchitecture and weights compatibility with an available spatio-temporalcomputer vision processor. The potential of the proposed taxonomy isdemonstrated for the specific case of point cloud classification by designingnovel compressed domain processors using the JPEG Pleno Point Cloud Codingstandard under development and adaptations of the PointGrid classifier.Experimental results show that the designed compressed domain point cloudclassification solutions can significantly outperform the spatial-temporaldomain classification benchmarks when applied to the decompressed data,containing coding artifacts, and even surpass their performance when applied tothe original uncompressed data.

Deep Learning-based Compressed Domain Multimedia for Man and Machine: A Taxonomy and Application to Point Cloud Classification