Depth Estimation and Blur Removal from a Single Out-of-focus Image

Saeed Anwar, Zeeshan Hayder and Fatih Porikli

Abstract

This paper presents a depth estimation method that leverages rich representations learned from cascaded convolutional and fully connected neural networks operating on a patch-pooled set of feature maps. Our method is very fast and it substantially improves depth accuracy over the state-of-the-art alternatives, and from this, we computationally reconstruct an all-focus image and achieve synthetic re-focusing, all from a single image. Our experiments on benchmark datasets such as Make3D and NYU-v2 demonstrate superior performance in comparison to other available depth estimation methods by reducing the root-mean-squared error by 57% & 46%, and blur removal methods by 0.36 dB & 0.72 dB in PSNR, respectively.

Session

Posters

Files

Paper (PDF)
Supplementary (PDF)

DOI

10.5244/C.31.113
https://dx.doi.org/10.5244/C.31.113

Citation

Bibtex

            @inproceedings{BMVC2017_113,
                title={Depth Estimation and Blur Removal from a Single Out-of-focus Image},
                author={Saeed Anwar, Zeeshan Hayder and Fatih Porikli},
                year={2017},
                month={September},
                pages={113.1-113.12},
                articleno={113},
                numpages={12},
                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.31.113},
                isbn={1-901725-60-X},
                url={https://dx.doi.org/10.5244/C.31.113}
            }