Learning confidence measures in the wild

Fabio Tosi, Matteo Poggi, Stefano Mattoccia, Alessio Tonioni and Luigi Di Stefano

Abstract

Confidence measures for stereo earned increasing popularity in most recent works concerning stereo, being effectively deployed to improve its accuracy. While most measures are obtained by processing cues from the cost volume, top-performing ones usually leverage on random-forests or CNNs to predict match reliability. Therefore, a proper amount of labeled data is required to effectively train such confidence measures. Being such ground-truth labels not always available in practical applications, in this paper we propose a methodology suited for training confidence measures in a self-supervised manner. Leveraging on a pool of properly selected conventional measures, we automatically detect a subset of very reliable pixels as well as a subset of erroneous samples from the output of a stereo algorithm. This strategy provides labels for training confidence measures based on machine-learning technique without ground-truth labels. Compared to state-of-the-art, our method is neither constrained to image sequences nor to image content.

Session

Posters

Files

Paper (PDF)
Supplementary (PDF)

DOI

10.5244/C.31.133
https://dx.doi.org/10.5244/C.31.133

Citation

Bibtex

            @inproceedings{BMVC2017_133,
                title={Learning confidence measures in the wild},
                author={Fabio Tosi, Matteo Poggi, Stefano Mattoccia, Alessio Tonioni and Luigi Di Stefano},
                year={2017},
                month={September},
                pages={133.1-133.13},
                articleno={133},
                numpages={13},
                booktitle={Proceedings of the British Machine Vision Conference (BMVC)},
                publisher={BMVA Press},
                editor={Tae-Kyun Kim, Stefanos Zafeiriou, Gabriel Brostow and Krystian Mikolajczyk},
                doi={10.5244/C.31.133},
                isbn={1-901725-60-X},
                url={https://dx.doi.org/10.5244/C.31.133}
            }