Activities per year
Abstract
In this work, we propose a learning-based method to denoise and refine disparity maps of a given stereo method. The proposed variational network arises naturally from unrolling the iterates of a proximal gradient method applied to a variational energy defined in a joint disparity, color, and confidence image space. Our method allows to learn a robust collaborative regularizer leveraging the joint statistics of the color image, the confidence map and the disparity map. Due to the variational structure of our method, the individual steps can be easily visualized, thus enabling interpretability of the method. We can therefore provide interesting insights into how our method refines and denoises disparity maps. The efficiency of our method is demonstrated by the publicly available stereo benchmarks Middlebury 2014 and Kitti 2015.
Original language | English |
---|---|
Title of host publication | German Conference on Pattern Recognition |
Pages | 3-17 |
Publication status | Published - 2019 |
Event | 41th German Conference on Pattern Recognition - Dortmund, Germany Duration: 10 Sept 2019 → 13 Sept 2019 http://gcpr2019.tu-dortmund.de |
Conference
Conference | 41th German Conference on Pattern Recognition |
---|---|
Abbreviated title | GCPR 2019 |
Country/Territory | Germany |
City | Dortmund |
Period | 10/09/19 → 13/09/19 |
Internet address |
Fingerprint
Dive into the research topics of 'Learned Collaborative Stereo Refinement'. Together they form a unique fingerprint.Activities
- 1 Talk at conference or symposium
-
Learned Collaborative Stereo Refinement
Knöbelreiter, P. (Speaker)
10 Sept 2019Activity: Talk or presentation › Talk at conference or symposium › Science to science
Prizes
-
GCPR Best Paper Award
Knöbelreiter, P. (Recipient) & Pock, T. (Recipient), 13 Sept 2019
Prize: Prizes / Medals / Awards