eduzhai > Applied Sciences > Engineering >

Monte Carlo Dropout Ensembles for Robust Illumination Estimation

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 7 pages

Abstract: Computational color constancy is a preprocessing step used in many camerasystems. The main aim is to discount the effect of the illumination on thecolors in the scene and restore the original colors of the objects. Recently,several deep learning-based approaches have been proposed to solve this problemand they often led to state-of-the-art performance in terms of average errors.However, for extreme samples, these methods fail and lead to high errors. Inthis paper, we address this limitation by proposing to aggregate different deeplearning methods according to their output uncertainty. We estimate therelative uncertainty of each approach using Monte Carlo dropout and the finalillumination estimate is obtained as the sum of the different model estimatesweighted by the log-inverse of their corresponding uncertainties. The proposedframework leads to state-of-the-art performance on INTEL-TAU dataset.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×