eduzhai > Applied Sciences > Engineering >

Image compression optimized for 3D reconstruction by utilizing deep neural networks

  • KanKan
  • (0) Download
  • 20210430
  • Save

... pages left unread,continue reading

Document pages: 12 pages

Abstract: Computer vision tasks are often expected to be executed on compressed images.Classical image compression standards like JPEG 2000 are widely used. However,they do not account for the specific end-task at hand. Motivated by works onrecurrent neural network (RNN)-based image compression and three-dimensional(3D) reconstruction, we propose unified network architectures to solve bothtasks jointly. These joint models provide image compression tailored for thespecific task of 3D reconstruction. Images compressed by our proposed models,yield 3D reconstruction performance superior as compared to using JPEG 2000compression. Our models significantly extend the range of compression rates forwhich 3D reconstruction is possible. We also show that this can be done highlyefficiently at almost no additional cost to obtain compression on top of thecomputation already required for performing the 3D reconstruction task.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...