arXiv reaDer
Deep Learning-based Compressed Domain Multimedia for Man and Machine: A Taxonomy and Application to Point Cloud Classification
In the current golden age of multimedia, human visualization is no longer the single main target, with the final consumer often being a machine which performs some processing or computer vision tasks. In both cases, deep learning plays a undamental role in extracting features from the multimedia representation data, usually producing a compressed representation referred to as latent representation. The increasing development and adoption of deep learning-based solutions in a wide area of multimedia applications have opened an exciting new vision where a common compressed multimedia representation is used 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 of coding artifacts are mitigated; and ii) reduced computational complexity, since prior decoding is not required. This paper proposes the first taxonomy for designing compressed domain computer vision solutions driven by the architecture and weights compatibility with an available spatio-temporal computer vision processor. The potential of the proposed taxonomy is demonstrated for the specific case of point cloud classification by designing novel compressed domain processors using the JPEG Pleno Point Cloud Coding standard under development and adaptations of the PointGrid classifier. Experimental results show that the designed compressed domain point cloud classification solutions can significantly outperform the spatial-temporal domain classification benchmarks when applied to the decompressed data, containing coding artifacts, and even surpass their performance when applied to the original uncompressed data.
updated: Fri Nov 17 2023 15:53:50 GMT+0000 (UTC)
published: Sat Oct 28 2023 23:38:30 GMT+0000 (UTC)
参考文献 (このサイトで利用可能なもの) / References (only if available on this site)
被参照文献 (このサイトで利用可能なものを新しい順に) / Citations (only if available on this site, in order of most recent)アソシエイト