Welcome to the new version of CaltechAUTHORS. Login is currently restricted to library staff. If you notice any issues, please email coda@library.caltech.edu
Published October 2020 | Published + Accepted Version
Book Section - Chapter Open

Fast Uncertainty Estimation for Deep Learning Based Optical Flow

Abstract

We present a novel approach to reduce the processing time required to derive the estimation uncertainty map in deep learning-based optical flow determination methods. Without uncertainty aware reasoning, the optical flow model, especially when it is used for mission critical fields such as robotics and aerospace, can cause catastrophic failures. Although several approaches such as the ones based on Bayesian neural networks have been proposed to handle this issue, they are computationally expensive. Thus, to speed up the processing time, our approach applies a generative model, which is trained by input images and an uncertainty map derived through a Bayesian approach. By using synthetically generated images of spacecraft, we demonstrate that the trained generative model can produce the uncertainty map 100∼700 times faster than the conventional uncertainty estimation method used for training the generative model itself. We also show that the quality of uncertainty map derived by the generative model is close to that of the original uncertainty map. By applying the proposed approach, the deep learning model operated in real-time can avoid disastrous failures by considering the uncertainty as well as achieving better performance removing uncertain portions of the prediction result.

Additional Information

© 2020 IEEE. The authors thank A. Rahmani, A. Santamaria-Navarro, and F. Y. Hadaegh for their technical guidance.

Attached Files

Published - 09340963.pdf

Accepted Version - IROS_2020_optical_flow_new_uncertainty__4_.pdf

Files

IROS_2020_optical_flow_new_uncertainty__4_.pdf
Files (2.6 MB)
Name Size Download all
md5:26fba3431060c8d11eccf0c18f3cc259
1.2 MB Preview Download
md5:13a98d7f453f4a2e89cbc0c2f1b60be6
1.3 MB Preview Download

Additional details

Created:
August 19, 2023
Modified:
October 20, 2023