Deep Depth from Defocus: How Can Defocus Blur Improve 3D Estimation Using Dense Neural Networks?

  • Marcela CarvalhoEmail author
  • Bertrand Le Saux
  • Pauline Trouvé-Peloux
  • Andrés Almansa
  • Frédéric Champagnat
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11129)


Depth estimation is critical interest for scene understanding and accurate 3D reconstruction. Most recent approaches with deep learning exploit geometrical structures of standard sharp images to predict depth maps. However, cameras can also produce images with defocus blur depending on the depth of the objects and camera settings. Hence, these features may represent an important hint for learning to predict depth. In this paper, we propose a full system for single-image depth prediction in the wild using depth-from-defocus and neural networks. We carry out thorough experiments real and simulated defocused images using a realistic model of blur variation with respect to depth. We also investigate the influence of blur on depth prediction observing model uncertainty with a Bayesian neural network approach. From these studies, we show that out-of-focus blur greatly improves the depth-prediction network performances. Furthermore, we transfer the ability learned on a synthetic, indoor dataset to real, indoor and outdoor images. For this purpose, we present a new dataset with real all-focus and defocused images from a DSLR camera, paired with ground truth depth maps obtained with an active 3D sensor for indoor scenes. The proposed approach is successfully validated on both this new dataset and standard ones as NYUv2 or Depth-in-the-Wild. Code and new datasets are available at


Depth from defocus Domain adaptation Depth estimation Single-image depth prediction 


  1. 1.
    Saxena, A., Sun, M., Ng, A.Y.: Make3D: learning 3D scene structure from a single still image. IEEE Trans. Pattern Anal. Mach. Intell. 31(5), 824–840 (2009)CrossRefGoogle Scholar
  2. 2.
    Calderero, F., Caselles, V.: Recovering relative depth from low-level features without explicit T-junction detection and interpretation. Int. J. Comput. Vis. 104(1), 38–68 (2013)MathSciNetCrossRefGoogle Scholar
  3. 3.
    Ng, R., Levoy, M., Brédif, M., Duval, G., Horowitz, M., Hanrahan, P.: Light field photography with a hand-held plenoptic camera. Computer Science Technical report (2005)Google Scholar
  4. 4.
    Eigen, D., Puhrsch, C., Fergus, R.: Depth map prediction from a single image using a multi-scale deep network. In: NIPS (2014)Google Scholar
  5. 5.
    Eigen, D., Fergus, R.: Predicting depth, surface normals and semantic labels with a common multi-scale convolutional architecture. In: ICCV (2015)Google Scholar
  6. 6.
    Li, B., Shen, C., Dai, Y., Van Den Hengel, A., He, M.: Depth and surface normal estimation from monocular images using regression on deep features and hierarchical CRFs (2015)Google Scholar
  7. 7.
    Wang, P., Shen, X., Lin, Z., Cohen, S., Price, B., Yuille, A.L.: Towards unified depth and semantic prediction from a single image. In: CVPR (2015)Google Scholar
  8. 8.
    Chakrabarti, A., Shao, J., Shakhnarovich, G.: Depth from a single image by harmonizing overcomplete local network predictions. In: NIPS (2016)Google Scholar
  9. 9.
    Ummenhofer, B., et al.: DeMoN: depth and motion network for learning monocular stereo. arXiv preprint arXiv:1612.02401 (2016)
  10. 10.
    Pentland, A.P.: A new sense for depth of field. IEEE Trans. PAMI 9(4), 523–531 (1987)CrossRefGoogle Scholar
  11. 11.
    Levin, A., Fergus, R., Durand, F., Freeman, W.T.: Image and depth from a conventional camera with a coded aperture. ACM Trans. Graph. 26, 70 (2007)CrossRefGoogle Scholar
  12. 12.
    Trouvé, P., Champagnat, F., Le Besnerais, G., Sabater, J., Avignon, T., Idier, J.: Passive depth estimation using chromatic aberration and a depth from defocus approach. Appl. Opt. 52(29), 7152–7164 (2013)CrossRefGoogle Scholar
  13. 13.
    Martinello, M., Favaro, P.: Single image blind deconvolution with higher-order texture statistics. In: Cremers, D., Magnor, M., Oswald, M.R., Zelnik-Manor, L. (eds.) Video Processing and Computational Video. LNCS, vol. 7082, pp. 124–151. Springer, Heidelberg (2011). Scholar
  14. 14.
    Sellent, A., Favaro, P.: Which side of the focal plane are you on? In: ICCP (2014)Google Scholar
  15. 15.
    Zhuo, S., Sim, T.: Defocus map estimation from a single image. Pattern Recognit. 44, 1852–1858 (2011)CrossRefGoogle Scholar
  16. 16.
    Carvalho, M., Saux, B.L., Trouvé-Peloux, P., Almansa, A., Champagnat, F.: On regression losses for deep depth estimation. In: ICIP (2018, to appear)Google Scholar
  17. 17.
    Chen, W., Fu, Z., Yang, D., Deng, J.: Single-image depth perception in the wild. In: NIPS, pp. 730–738 (2016)Google Scholar
  18. 18.
    Saxena, A., Chung, S.H., Ng, A.Y.: Learning depth from single monocular images. In: NIPS (2006)Google Scholar
  19. 19.
    Cao, Y., Wu, Z., Shen, C.: Estimating depth from monocular images as classification using deep fully convolutional residual networks. IEEE Trans. Circuits Syst. Video Technol. 28(11), 3174–3182 (2017)CrossRefGoogle Scholar
  20. 20.
    He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: CVPR, pp. 770–778 (2016)Google Scholar
  21. 21.
    Xu, D., Ricci, E., Ouyang, W., Wang, X., Sebe, N.: Multi-scale continuous CRFs as sequential deep networks for monocular depth estimation. arXiv preprint arXiv:1704.02157 (2017)
  22. 22.
    Laina, I., Rupprecht, C., Belagiannis, V., Tombari, F., Navab, N.: Deeper depth prediction with fully convolutional residual networks. In: 2016 Fourth International Conference on 3D Vision (3DV), pp. 239–248. IEEE (2016)Google Scholar
  23. 23.
    Jung, H., Kim, Y., Min, D., Oh, C., Sohn, K.: Depth prediction from a single image with conditional adversarial networks. In: ICIP (2017)Google Scholar
  24. 24.
    Goodfellow, I., et al.: Generative adversarial nets. In: NIPS (2014)Google Scholar
  25. 25.
    Godard, C., Mac Aodha, O., Brostow, G.J.: Unsupervised monocular depth estimation with left-right consistency. arXiv preprint arXiv:1609.03677 (2016)
  26. 26.
    Garg, R., B.G., V.K., Carneiro, G., Reid, I.: Unsupervised CNN for single view depth estimation: geometry to the rescue. In: Leibe, B., Matas, J., Sebe, N., Welling, M. (eds.) ECCV 2016. LNCS, vol. 9912, pp. 740–756. Springer, Cham (2016). Scholar
  27. 27.
    Kendall, A., Gal, Y.: What uncertainties do we need in Bayesian deep learning for computer vision? arXiv preprint arXiv:1703.04977 (2017)
  28. 28.
    Huang, G., Liu, Z., van der Maaten, L., Weinberger, K.Q.: Densely connected convolutional networks. In: CVPR (2017)Google Scholar
  29. 29.
    Levin, A., Weiss, Y., Durand, F., Freeman, W.: Understanding and evaluating blind deconvolution algorithms. In: CVPR, pp. 1–8 (2009)Google Scholar
  30. 30.
    Veeraraghavan, A., Raskar, R., Agrawal, A., Mohan, A., Tumblin, J.: Dappled photography: mask enhanced cameras for heterodyned light fields and coded aperture refocusing. ACM Trans. Graph. 26, 69 (2007)CrossRefGoogle Scholar
  31. 31.
    Chakrabarti, A., Zickler, T.: Depth and deblurring from a spectrally-varying depth-of-field. In: Fitzgibbon, A., Lazebnik, S., Perona, P., Sato, Y., Schmid, C. (eds.) ECCV 2012. LNCS, vol. 7576, pp. 648–661. Springer, Heidelberg (2012). Scholar
  32. 32.
    Hazirbas, C., Leal-Taixé, L., Cremers, D.: Deep depth from focus. arxiv preprint arXiv:1704.01085, April 2017
  33. 33.
    Guichard, F., Nguyen, H.P., Tessières, R., Pyanet, M., Tarchouna, I., Cao, F.: Extended depth-of-field using sharpness transport across color channels. In: Rodricks, B.G., Süsstrunk, S.E. (eds.) IS&T/SPIE Electronic Imaging, International Society for Optics and Photonics, pp. 72500N–72500N-12, January 2009Google Scholar
  34. 34.
    Delbracio, M., Musé, P., Almansa, A., Morel, J.: The non-parametric sub-pixel local point spread function estimation is a well posed problem. Int. J. Comput. Vis. 96, 175–194 (2012)MathSciNetCrossRefGoogle Scholar
  35. 35.
    Silberman, N., Hoiem, D., Kohli, P., Fergus, R.: Indoor segmentation and support inference from RGBD images. In: Fitzgibbon, A., Lazebnik, S., Perona, P., Sato, Y., Schmid, C. (eds.) ECCV 2012. LNCS, vol. 7576, pp. 746–760. Springer, Heidelberg (2012). Scholar
  36. 36.
    Srinivasan, P.P., Garg, R., Wadhwa, N., Ng, R., Barron, J.T.: Aperture supervision for monocular depth estimation. arXiv preprint arXiv:1711.07933 (2016)
  37. 37.
    Anwar, S., Hayder, Z., Porikli, F.: Depth estimation and blur removal from a single out-of-focus image. In: BMVC (2017)Google Scholar
  38. 38.
    Ronneberger, O., Fischer, P., Brox, T.: U-net: convolutional networks for biomedical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015. LNCS, vol. 9351, pp. 234–241. Springer, Cham (2015). Scholar
  39. 39.
    Srivastava, N., Hinton, G.E., Krizhevsky, A., Sutskever, I., Salakhutdinov, R.: Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Res. 15(1), 1929–1958 (2014)MathSciNetzbMATHGoogle Scholar
  40. 40.
    Hasinoff, S.W., Kutulakos, K.N.: A layer-based restoration framework for variable-aperture photography. In: 2007 IEEE 11th International Conference on Computer Vision, pp. 1–8, October 2007Google Scholar
  41. 41.
    Trouvé, P., Champagnat, F., Le Besnerais, G., Idier, J.: Single image local blur identification. In: IEEE ICIP (2011)Google Scholar
  42. 42.
    Kendall, A., Badrinarayanan, V., Cipolla, R.: Bayesian SegNet: model uncertainty in deep convolutional encoder-decoder architectures for scene understanding. arXiv preprint arXiv:1511.02680 (2015)
  43. 43.
    Gal, Y., Ghahramani, Z.: Dropout as a Bayesian approximation: representing model uncertainty in deep learning. In: ICML, pp. 1050–1059 (2016)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Marcela Carvalho
    • 1
    Email author
  • Bertrand Le Saux
    • 1
  • Pauline Trouvé-Peloux
    • 1
  • Andrés Almansa
    • 2
  • Frédéric Champagnat
    • 2
  1. 1.DTIS, ONERA, Université Paris-SaclayPalaiseauFrance
  2. 2.Université Paris DescartesParisFrance

Personalised recommendations