Current Year

Article
Beigpour, S., Shekhar, S., Mansouryar, M., Myszkowski, K., and Seidel, H.-P. Light-Field Appearance Editing Based on Intrinsic Decomposition. Journal of Perceptual Imaging.
(Accepted/in press)
Export
BibTeX
@article{Beigpour2018, TITLE = {Light-Field Appearance Editing Based on Intrinsic Decomposition}, AUTHOR = {Beigpour, Shida and Shekhar, Sumit and Mansouryar, Mohsen and Myszkowski, Karol and Seidel, Hans-Peter}, LANGUAGE = {eng}, YEAR = {2018}, PUBLREMARK = {Accepted}, JOURNAL = {Journal of Perceptual Imaging}, }
Endnote
%0 Journal Article %A Beigpour, Shida %A Shekhar, Sumit %A Mansouryar, Mohsen %A Myszkowski, Karol %A Seidel, Hans-Peter %+ External Organizations Computer Graphics, MPI for Informatics, Max Planck Society External Organizations Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society %T Light-Field Appearance Editing Based on Intrinsic Decomposition : %G eng %U http://hdl.handle.net/21.11116/0000-0001-5F88-C %D 2018 %J Journal of Perceptual Imaging %O JPI
Chen, R., Gotsman, C., and Hormann, K. 2018. Efficient Path Generation with Reduced Coordinates. Computer Graphics Forum (Proc. Eurographics Symposium on Geometric Processing 2018)37, 5.
Export
BibTeX
@article{ChenSGP2018, TITLE = {Efficient Path Generation with Reduced Coordinates}, AUTHOR = {Chen, Renjie and Gotsman, Craig and Hormann, Kai}, LANGUAGE = {eng}, ISSN = {0167-7055}, DOI = {10.1111/cgf.13489}, PUBLISHER = {Wiley-Blackwell}, ADDRESS = {Chichester}, YEAR = {2018}, DATE = {2018}, JOURNAL = {Computer Graphics Forum (Proc. Eurographics Symposium on Geometric Processing)}, VOLUME = {37}, NUMBER = {5}, PAGES = {37--48}, BOOKTITLE = {Symposium on Geometry Processing 2018 (Eurographics Symposium on Geometric Processing 2018)}, EDITOR = {Ju, Tao and Vaxman, Amir}, }
Endnote
%0 Journal Article %A Chen, Renjie %A Gotsman, Craig %A Hormann, Kai %+ Computer Graphics, MPI for Informatics, Max Planck Society External Organizations External Organizations %T Efficient Path Generation with Reduced Coordinates : %G eng %U http://hdl.handle.net/21.11116/0000-0001-E6D6-A %R 10.1111/cgf.13489 %7 2018 %D 2018 %J Computer Graphics Forum %V 37 %N 5 %& 37 %P 37 - 48 %I Wiley-Blackwell %C Chichester %@ false %B Symposium on Geometry Processing 2018 %O Paris, France, July 7 – 11, 2018 SGP 2018 Eurographics Symposium on Geometric Processing 2018
Du, X., Liu, X., Yan, D.-M., Jiang, C., Ye, J., and Zhang, H. 2018. Field-Aligned Isotropic Surface Remeshing. Computer Graphics Forum37, 6.
Export
BibTeX
@article{DuCGF2018, TITLE = {Field-Aligned Isotropic Surface Remeshing}, AUTHOR = {Du, Xingyi and Liu, Xiaohan and Yan, Dong-Ming and Jiang, Caigui and Ye, Juntao and Zhang, Hui}, LANGUAGE = {eng}, ISSN = {0167-7055}, DOI = {10.1111/cgf.13329}, PUBLISHER = {Blackwell-Wiley}, ADDRESS = {Oxford}, YEAR = {2018}, DATE = {2018}, JOURNAL = {Computer Graphics Forum}, VOLUME = {37}, NUMBER = {6}, PAGES = {343--357}, }
Endnote
%0 Journal Article %A Du, Xingyi %A Liu, Xiaohan %A Yan, Dong-Ming %A Jiang, Caigui %A Ye, Juntao %A Zhang, Hui %+ External Organizations External Organizations External Organizations Computer Graphics, MPI for Informatics, Max Planck Society External Organizations External Organizations %T Field-Aligned Isotropic Surface Remeshing : %G eng %U http://hdl.handle.net/21.11116/0000-0001-E209-6 %R 10.1111/cgf.13329 %7 2018 %D 2018 %J Computer Graphics Forum %O Computer Graphics Forum : journal of the European Association for Computer Graphics Comput. Graph. Forum %V 37 %N 6 %& 343 %P 343 - 357 %I Blackwell-Wiley %C Oxford %@ false
Leimkühler, T., Kellnhofer, P., Ritschel, T., Myszkowski, K., and Seidel, H.-P. 2018. Perceptual Real-Time 2D-to-3D Conversion Using Cue Fusion. IEEE Transactions on Visualization and Computer Graphics24, 6.
Export
BibTeX
@article{Leimkuehler2018, TITLE = {Perceptual real-time {2D}-to-{3D} conversion using cue fusion}, AUTHOR = {Leimk{\"u}hler, Thomas and Kellnhofer, Petr and Ritschel, Tobias and Myszkowski, Karol and Seidel, Hans-Peter}, LANGUAGE = {eng}, ISSN = {1077-2626}, DOI = {10.1109/TVCG.2017.2703612}, PUBLISHER = {IEEE Computer Society}, ADDRESS = {New York, NY}, YEAR = {2018}, DATE = {2018}, JOURNAL = {IEEE Transactions on Visualization and Computer Graphics}, VOLUME = {24}, NUMBER = {6}, PAGES = {2037--2050}, }
Endnote
%0 Journal Article %A Leimkühler, Thomas %A Kellnhofer, Petr %A Ritschel, Tobias %A Myszkowski, Karol %A Seidel, Hans-Peter %+ Computer Graphics, MPI for Informatics, Max Planck Society External Organizations External Organizations Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society %T Perceptual Real-Time 2D-to-3D Conversion Using Cue Fusion : %G eng %U http://hdl.handle.net/21.11116/0000-0001-409A-9 %R 10.1109/TVCG.2017.2703612 %7 2018 %D 2018 %J IEEE Transactions on Visualization and Computer Graphics %V 24 %N 6 %& 2037 %P 2037 - 2050 %I IEEE Computer Society %C New York, NY %@ false
Thies, J., Zollhöfer, M., Stamminger, M., Theobalt, C., and Nießner, M. 2018. FaceVR: Real-Time Facial Reenactment and Eye Gaze Control in Virtual Reality. ACM Transactions on Graphics37, 2.
Export
BibTeX
@article{Thies_ToG2018, TITLE = {{FaceVR}: Real-Time Facial Reenactment and Eye Gaze Control in Virtual Reality}, AUTHOR = {Thies, Justus and Zollh{\"o}fer, Michael and Stamminger, Marc and Theobalt, Christian and Nie{\ss}ner, Matthias}, LANGUAGE = {eng}, ISSN = {0730-0301}, DOI = {10.1145/3182644}, PUBLISHER = {ACM}, ADDRESS = {New York, NY}, YEAR = {2018}, DATE = {2018}, JOURNAL = {ACM Transactions on Graphics}, VOLUME = {37}, NUMBER = {2}, EID = {25}, }
Endnote
%0 Journal Article %A Thies, Justus %A Zollhöfer, Michael %A Stamminger, Marc %A Theobalt, Christian %A Nießner, Matthias %+ External Organizations External Organizations External Organizations Computer Graphics, MPI for Informatics, Max Planck Society External Organizations %T FaceVR: Real-Time Facial Reenactment and Eye Gaze Control in Virtual Reality : %G eng %U http://hdl.handle.net/21.11116/0000-0001-E215-8 %R 10.1145/3182644 %7 2018 %D 2018 %J ACM Transactions on Graphics %V 37 %N 2 %Z sequence number: 25 %I ACM %C New York, NY %@ false
Wolski, K., Giunchi, D., Ye, N., et al. Dataset and Metrics for Predicting Local Visible Differences. ACM Transactions on Graphics.
(Accepted/in press)
Export
BibTeX
@article{wolski2018dataset, TITLE = {Dataset and Metrics for Predicting Local Visible Differences}, AUTHOR = {Wolski, Krzysztof and Giunchi, Daniele and Ye, Nanyang and Didyk, Piotr and Myszkowski, Karol and Mantiuk, Rados{\textbackslash}l{\textbraceleft}{\textbraceright}aw and Seidel, Hans-Peter and Steed, Anthony and Mantiuk, Rafa{\l} K.}, ISSN = {0730-0301}, PUBLISHER = {Association for Computing Machinery}, ADDRESS = {New York, NY}, YEAR = {2018}, PUBLREMARK = {Accepted}, JOURNAL = {ACM Transactions on Graphics}, }
Endnote
%0 Journal Article %A Wolski, Krzysztof %A Giunchi, Daniele %A Ye, Nanyang %A Didyk, Piotr %A Myszkowski, Karol %A Mantiuk, Rados\l{}aw %A Seidel, Hans-Peter %A Steed, Anthony %A Mantiuk, Rafał K. %+ Computer Graphics, MPI for Informatics, Max Planck Society External Organizations External Organizations Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society External Organizations Computer Graphics, MPI for Informatics, Max Planck Society External Organizations External Organizations %T Dataset and Metrics for Predicting Local Visible Differences : %U http://hdl.handle.net/21.11116/0000-0001-5F75-2 %D 2018 %J ACM Transactions on Graphics %I Association for Computing Machinery %C New York, NY %@ false
Xu, W., Chatterjee, A., Zollhöfer, M., et al. 2018a. MonoPerfCap: Human Performance Capture from Monocular Video. ACM Transactions on Graphics37, 2.
Abstract
We present the first marker-less approach for temporally coherent 3D performance capture of a human with general clothing from monocular video. Our approach reconstructs articulated human skeleton motion as well as medium-scale non-rigid surface deformations in general scenes. Human performance capture is a challenging problem due to the large range of articulation, potentially fast motion, and considerable non-rigid deformations, even from multi-view data. Reconstruction from monocular video alone is drastically more challenging, since strong occlusions and the inherent depth ambiguity lead to a highly ill-posed reconstruction problem. We tackle these challenges by a novel approach that employs sparse 2D and 3D human pose detections from a convolutional neural network using a batch-based pose estimation strategy. Joint recovery of per-batch motion allows to resolve the ambiguities of the monocular reconstruction problem based on a low dimensional trajectory subspace. In addition, we propose refinement of the surface geometry based on fully automatically extracted silhouettes to enable medium-scale non-rigid alignment. We demonstrate state-of-the-art performance capture results that enable exciting applications such as video editing and free viewpoint video, previously infeasible from monocular video. Our qualitative and quantitative evaluation demonstrates that our approach significantly outperforms previous monocular methods in terms of accuracy, robustness and scene complexity that can be handled.
Export
BibTeX
@article{Xu_ToG2018, TITLE = {{MonoPerfCap}: Human Performance Capture from Monocular Video}, AUTHOR = {Xu, Weipeng and Chatterjee, Avishek and Zollh{\"o}fer, Michael and Rhodin, Helge and Mehta, Dushyant and Seidel, Hans-Peter and Theobalt, Christian}, LANGUAGE = {eng}, ISSN = {0730-0301}, DOI = {10.1145/3181973}, PUBLISHER = {ACM}, ADDRESS = {New York, NY}, YEAR = {2018}, DATE = {2018}, ABSTRACT = {We present the first marker-less approach for temporally coherent 3D performance capture of a human with general clothing from monocular video. Our approach reconstructs articulated human skeleton motion as well as medium-scale non-rigid surface deformations in general scenes. Human performance capture is a challenging problem due to the large range of articulation, potentially fast motion, and considerable non-rigid deformations, even from multi-view data. Reconstruction from monocular video alone is drastically more challenging, since strong occlusions and the inherent depth ambiguity lead to a highly ill-posed reconstruction problem. We tackle these challenges by a novel approach that employs sparse 2D and 3D human pose detections from a convolutional neural network using a batch-based pose estimation strategy. Joint recovery of per-batch motion allows to resolve the ambiguities of the monocular reconstruction problem based on a low dimensional trajectory subspace. In addition, we propose refinement of the surface geometry based on fully automatically extracted silhouettes to enable medium-scale non-rigid alignment. We demonstrate state-of-the-art performance capture results that enable exciting applications such as video editing and free viewpoint video, previously infeasible from monocular video. Our qualitative and quantitative evaluation demonstrates that our approach significantly outperforms previous monocular methods in terms of accuracy, robustness and scene complexity that can be handled.}, JOURNAL = {ACM Transactions on Graphics}, VOLUME = {37}, NUMBER = {2}, EID = {27}, }
Endnote
%0 Journal Article %A Xu, Weipeng %A Chatterjee, Avishek %A Zollhöfer, Michael %A Rhodin, Helge %A Mehta, Dushyant %A Seidel, Hans-Peter %A Theobalt, Christian %+ Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society External Organizations Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society %T MonoPerfCap: Human Performance Capture from Monocular Video : %G eng %U http://hdl.handle.net/21.11116/0000-0001-E20E-1 %R 10.1145/3181973 %7 2017 %D 2018 %X We present the first marker-less approach for temporally coherent 3D performance capture of a human with general clothing from monocular video. Our approach reconstructs articulated human skeleton motion as well as medium-scale non-rigid surface deformations in general scenes. Human performance capture is a challenging problem due to the large range of articulation, potentially fast motion, and considerable non-rigid deformations, even from multi-view data. Reconstruction from monocular video alone is drastically more challenging, since strong occlusions and the inherent depth ambiguity lead to a highly ill-posed reconstruction problem. We tackle these challenges by a novel approach that employs sparse 2D and 3D human pose detections from a convolutional neural network using a batch-based pose estimation strategy. Joint recovery of per-batch motion allows to resolve the ambiguities of the monocular reconstruction problem based on a low dimensional trajectory subspace. In addition, we propose refinement of the surface geometry based on fully automatically extracted silhouettes to enable medium-scale non-rigid alignment. We demonstrate state-of-the-art performance capture results that enable exciting applications such as video editing and free viewpoint video, previously infeasible from monocular video. Our qualitative and quantitative evaluation demonstrates that our approach significantly outperforms previous monocular methods in terms of accuracy, robustness and scene complexity that can be handled. %K Computer Science, Computer Vision and Pattern Recognition, cs.CV,Computer Science, Graphics, cs.GR %J ACM Transactions on Graphics %V 37 %N 2 %Z sequence number: 27 %I ACM %C New York, NY %@ false
Zollhöfer, M., Stotko, P., Görlitz, A., et al. 2018a. State of the Art on 3D Reconstruction with RGB-D Cameras. Computer Graphics Forum (Proc. EUROGRAPHICS 2018)37, 2.
Export
BibTeX
@article{Zollhoefer_EG2018STAR, TITLE = {State of the Art on {3D} Reconstruction with {RGB}-{D} Cameras}, AUTHOR = {Zollh{\"o}fer, Michael and Stotko, Patrick and G{\"o}rlitz, Andreas and Theobalt, Christian and Nie{\ss}ner, Matthias and Klein, Reinhard and Kolb, Andreas}, LANGUAGE = {eng}, ISSN = {0167-7055}, DOI = {10.1111/cgf.13386}, PUBLISHER = {Blackwell-Wiley}, ADDRESS = {Oxford}, YEAR = {2018}, DATE = {2018}, JOURNAL = {Computer Graphics Forum (Proc. EUROGRAPHICS)}, VOLUME = {37}, NUMBER = {2}, PAGES = {625--652}, BOOKTITLE = {EUROGRAPHICS 2018 STAR -- State of The Art Reports}, EDITOR = {Hildebrandt, Klaus and Theobalt, Christian}, }
Endnote
%0 Journal Article %A Zollhöfer, Michael %A Stotko, Patrick %A Görlitz, Andreas %A Theobalt, Christian %A Nießner, Matthias %A Klein, Reinhard %A Kolb, Andreas %+ Computer Graphics, MPI for Informatics, Max Planck Society External Organizations External Organizations Computer Graphics, MPI for Informatics, Max Planck Society External Organizations External Organizations External Organizations %T State of the Art on 3D Reconstruction with RGB-D Cameras : %G eng %U http://hdl.handle.net/21.11116/0000-0001-6917-0 %R 10.1111/cgf.13386 %7 2018 %D 2018 %J Computer Graphics Forum %O Computer Graphics Forum : journal of the European Association for Computer Graphics Comput. Graph. Forum %V 37 %N 2 %& 625 %P 625 - 652 %I Blackwell-Wiley %C Oxford %@ false %B EUROGRAPHICS 2018 STAR – State of The Art Reports %O EUROGRAPHICS 2018 EG 2018 The European Association for Computer Graphics 39th Annual Conference ; Delft, The Netherlands, April 16th - 20th, 2018
Zollhöfer, M., Thies, J., Garrido, P., et al. 2018b. State of the Art on Monocular 3D Face Reconstruction, Tracking, and Applications. Computer Graphics Forum (Proc. EUROGRAPHICS 2018)37, 2.
Export
BibTeX
@article{Zollhoefer_EG2018STARb, TITLE = {State of the Art on Monocular {3D} Face Reconstruction, Tracking, and Applications}, AUTHOR = {Zollh{\"o}fer, Michael and Thies, Justus and Garrido, Pablo and Bradley, D. and Beeler, T. and P{\'e}rez, P. and Stamminger, Marc and Nie{\ss}ner, Matthias and Theobalt, Christian}, LANGUAGE = {eng}, ISSN = {0167-7055}, DOI = {10.1111/cgf.13382}, PUBLISHER = {Blackwell-Wiley}, ADDRESS = {Oxford}, YEAR = {2018}, DATE = {2018}, JOURNAL = {Computer Graphics Forum (Proc. EUROGRAPHICS)}, VOLUME = {37}, NUMBER = {2}, PAGES = {523--550}, BOOKTITLE = {EUROGRAPHICS 2018 STAR -- State of The Art Reports}, EDITOR = {Hildebrandt, Klaus and Theobalt, Christian}, }
Endnote
%0 Journal Article %A Zollhöfer, Michael %A Thies, Justus %A Garrido, Pablo %A Bradley, D. %A Beeler, T. %A Pérez, P. %A Stamminger, Marc %A Nießner, Matthias %A Theobalt, Christian %+ Computer Graphics, MPI for Informatics, Max Planck Society External Organizations Computer Graphics, MPI for Informatics, Max Planck Society External Organizations External Organizations External Organizations External Organizations External Organizations Computer Graphics, MPI for Informatics, Max Planck Society %T State of the Art on Monocular 3D Face Reconstruction, Tracking, and Applications : %G eng %U http://hdl.handle.net/21.11116/0000-0001-691A-D %R 10.1111/cgf.13382 %7 2018 %D 2018 %J Computer Graphics Forum %O Computer Graphics Forum : journal of the European Association for Computer Graphics Comput. Graph. Forum %V 37 %N 2 %& 523 %P 523 - 550 %I Blackwell-Wiley %C Oxford %@ false %B EUROGRAPHICS 2018 STAR – State of The Art Reports %O EUROGRAPHICS 2018 EG 2018 The European Association for Computer Graphics 39th Annual Conference ; Delft, The Netherlands, April 16th - 20th, 2018
Conference Paper
Alldieck, T., Magnor, M.A., Xu, W., Theobalt, C., and Pons-Moll, G. Video Based Reconstruction of 3D People Models. 31st IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2018).
(Accepted/in press)
Export
BibTeX
@inproceedings{alldieck2018video, TITLE = {Video Based Reconstruction of {3D} People Models}, AUTHOR = {Alldieck, Thiemo and Magnor, Marcus A. and Xu, Weipeng and Theobalt, Christian and Pons-Moll, Gerard}, LANGUAGE = {eng}, YEAR = {2018}, PUBLREMARK = {Accepted}, BOOKTITLE = {31st IEEE Conference on Computer Vision and Pattern Recognition (CVPR 2018)}, ADDRESS = {Salt Lake City, UT, USA}, }
Endnote
%0 Conference Proceedings %A Alldieck, Thiemo %A Magnor, Marcus A. %A Xu, Weipeng %A Theobalt, Christian %A Pons-Moll, Gerard %+ External Organizations External Organizations External Organizations Computer Graphics, MPI for Informatics, Max Planck Society Computer Vision and Multimodal Computing, MPI for Informatics, Max Planck Society %T Video Based Reconstruction of 3D People Models : %G eng %U http://hdl.handle.net/21.11116/0000-0001-1E24-6 %D 2018 %B 31st IEEE Conference on Computer Vision and Pattern Recognition %Z date of event: 2018-06-18 - 2018-06-22 %C Salt Lake City, UT, USA %B 31st IEEE Conference on Computer Vision and Pattern Recognition
Myszkowski, K., Tursun, O.T., Kellnhofer, P., et al. Perceptual Display: Apparent Enhancement of Scene Detail and Depth. (Proc. HVEI 2018), SPIE/IS&T.
(Accepted/in press, Keynote Talk)
Abstract
Predicting human visual perception of image differences has several applications such as compression, rendering, editing and retargeting. Current approaches however, ignore the fact that the human visual system compensates for geometric transformations, e.g. we see that an image and a rotated copy are identical. Instead, they will report a large, false-positive difference. At the same time, if the transformations become too strong or too spatially incoherent, comparing two images indeed gets increasingly difficult. Between these two extremes, we propose a system to quantify the effect of transformations, not only on the perception of image differences, but also on saliency and motion parallax. To this end, we first fit local homographies to a given optical flow field and then convert this field into a field of elementary transformations such as translation, rotation, scaling, and perspective. We conduct a perceptual experiment quantifying the increase of difficulty when compensating for elementary transformations. Transformation entropy is proposed as a novel measure of complexity in a flow field. This representation is then used for applications, such as comparison of non-aligned images, where transformations cause threshold elevation, and detection of salient transformations.
Export
BibTeX
@inproceedings{Myszkowski2018Perceptual, TITLE = {Perceptual Display: Apparent Enhancement of Scene Detail and Depth}, AUTHOR = {Myszkowski, Karol and Tursun, Okan Tarhan and Kellnhofer, Petr and Templin, Krzysztof and Arabadzhiyska, Elena and Didyk, Piotr and Seidel, Hans-Peter}, LANGUAGE = {eng}, PUBLISHER = {SPIE/IS\&T}, YEAR = {2018}, PUBLREMARK = {Accepted}, ABSTRACT = {Predicting human visual perception of image differences has several applications such as compression, rendering, editing and retargeting. Current approaches however, ignore the fact that the human visual system compensates for geometric transformations, e.g. we see that an image and a rotated copy are identical. Instead, they will report a large, false-positive difference. At the same time, if the transformations become too strong or too spatially incoherent, comparing two images indeed gets increasingly difficult. Between these two extremes, we propose a system to quantify the effect of transformations, not only on the perception of image differences, but also on saliency and motion parallax. To this end, we first fit local homographies to a given optical flow field and then convert this field into a field of elementary transformations such as translation, rotation, scaling, and perspective. We conduct a perceptual experiment quantifying the increase of difficulty when compensating for elementary transformations. Transformation entropy is proposed as a novel measure of complexity in a flow field. This representation is then used for applications, such as comparison of non-aligned images, where transformations cause threshold elevation, and detection of salient transformations.}, BOOKTITLE = {Human Vision and Electronic Imaging (HVEI 2018)}, JOURNAL = {(Proc. HVEI)}, ADDRESS = {San Francisco, CA, USA}, }
Endnote
%0 Conference Proceedings %A Myszkowski, Karol %A Tursun, Okan Tarhan %A Kellnhofer, Petr %A Templin, Krzysztof %A Arabadzhiyska, Elena %A Didyk, Piotr %A Seidel, Hans-Peter %+ Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society %T Perceptual Display: Apparent Enhancement of Scene Detail and Depth : %G eng %U http://hdl.handle.net/21.11116/0000-0001-5F64-5 %D 2018 %B Human Vision and Electronic Imaging %Z date of event: 2018-01-28 - 2018-02-02 %C San Francisco, CA, USA %X Predicting human visual perception of image differences has several applications such as compression, rendering, editing and retargeting. Current approaches however, ignore the fact that the human visual system compensates for geometric transformations, e.g. we see that an image and a rotated copy are identical. Instead, they will report a large, false-positive difference. At the same time, if the transformations become too strong or too spatially incoherent, comparing two images indeed gets increasingly difficult. Between these two extremes, we propose a system to quantify the effect of transformations, not only on the perception of image differences, but also on saliency and motion parallax. To this end, we first fit local homographies to a given optical flow field and then convert this field into a field of elementary transformations such as translation, rotation, scaling, and perspective. We conduct a perceptual experiment quantifying the increase of difficulty when compensating for elementary transformations. Transformation entropy is proposed as a novel measure of complexity in a flow field. This representation is then used for applications, such as comparison of non-aligned images, where transformations cause threshold elevation, and detection of salient transformations. %B Human Vision and Electronic Imaging %I SPIE/IS&T
Robertini, N., Bernard, F., Xu, W., and Theobalt, C. 2018. Illumination-Invariant Robust Multiview 3D Human Motion Capture. 2018 IEEE Winter Conference on Applications of Computer Vision (WACV 2018), IEEE.
Export
BibTeX
@inproceedings{Robertini_WACV2018, TITLE = {Illumination-Invariant Robust Multiview {3D} Human Motion Capture}, AUTHOR = {Robertini, Nadia and Bernard, Florian and Xu, Weipeng and Theobalt, Christian}, LANGUAGE = {eng}, ISBN = {978-1-5386-4886-5}, DOI = {10.1109/WACV.2018.00185}, PUBLISHER = {IEEE}, YEAR = {2017}, DATE = {2018}, BOOKTITLE = {2018 IEEE Winter Conference on Applications of Computer Vision (WACV 2018)}, PAGES = {1661--1670}, ADDRESS = {Lake Tahoe, NV}, }
Endnote
%0 Conference Proceedings %A Robertini, Nadia %A Bernard, Florian %A Xu, Weipeng %A Theobalt, Christian %+ Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society %T Illumination-Invariant Robust Multiview 3D Human Motion Capture : %G eng %U http://hdl.handle.net/21.11116/0000-0001-A474-3 %R 10.1109/WACV.2018.00185 %D 2018 %B IEEE Winter Conference on Applications of Computer Vision %Z date of event: 2017-03-12 - 2017-03-15 %C Lake Tahoe, NV %B 2018 IEEE Winter Conference on Applications of Computer Vision %P 1661 - 1670 %I IEEE %@ 978-1-5386-4886-5
Paper
Alldieck, T., Magnor, M.A., Xu, W., Theobalt, C., and Pons-Moll, G. 2018. Video Based Reconstruction of 3D People Models. http://arxiv.org/abs/1803.04758.
(arXiv: 1803.04758)
Abstract
This paper describes how to obtain accurate 3D body models and texture of arbitrary people from a single, monocular video in which a person is moving. Based on a parametric body model, we present a robust processing pipeline achieving 3D model fits with 5mm accuracy also for clothed people. Our main contribution is a method to nonrigidly deform the silhouette cones corresponding to the dynamic human silhouettes, resulting in a visual hull in a common reference frame that enables surface reconstruction. This enables efficient estimation of a consensus 3D shape, texture and implanted animation skeleton based on a large number of frames. We present evaluation results for a number of test subjects and analyze overall performance. Requiring only a smartphone or webcam, our method enables everyone to create their own fully animatable digital double, e.g., for social VR applications or virtual try-on for online fashion shopping.
Export
BibTeX
@online{Alldieck_arXiv1803.04758, TITLE = {Video Based Reconstruction of {3D} People Models}, AUTHOR = {Alldieck, Thiemo and Magnor, Marcus A. and Xu, Weipeng and Theobalt, Christian and Pons-Moll, Gerard}, LANGUAGE = {eng}, URL = {http://arxiv.org/abs/1803.04758}, EPRINT = {1803.04758}, EPRINTTYPE = {arXiv}, YEAR = {2018}, ABSTRACT = {This paper describes how to obtain accurate 3D body models and texture of arbitrary people from a single, monocular video in which a person is moving. Based on a parametric body model, we present a robust processing pipeline achieving 3D model fits with 5mm accuracy also for clothed people. Our main contribution is a method to nonrigidly deform the silhouette cones corresponding to the dynamic human silhouettes, resulting in a visual hull in a common reference frame that enables surface reconstruction. This enables efficient estimation of a consensus 3D shape, texture and implanted animation skeleton based on a large number of frames. We present evaluation results for a number of test subjects and analyze overall performance. Requiring only a smartphone or webcam, our method enables everyone to create their own fully animatable digital double, e.g., for social VR applications or virtual try-on for online fashion shopping.}, }
Endnote
%0 Report %A Alldieck, Thiemo %A Magnor, Marcus A. %A Xu, Weipeng %A Theobalt, Christian %A Pons-Moll, Gerard %+ External Organizations External Organizations Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society D2 Extern %T Video Based Reconstruction of 3D People Models : %G eng %U http://hdl.handle.net/21.11116/0000-0001-40CD-0 %U http://arxiv.org/abs/1803.04758 %D 2018 %X This paper describes how to obtain accurate 3D body models and texture of arbitrary people from a single, monocular video in which a person is moving. Based on a parametric body model, we present a robust processing pipeline achieving 3D model fits with 5mm accuracy also for clothed people. Our main contribution is a method to nonrigidly deform the silhouette cones corresponding to the dynamic human silhouettes, resulting in a visual hull in a common reference frame that enables surface reconstruction. This enables efficient estimation of a consensus 3D shape, texture and implanted animation skeleton based on a large number of frames. We present evaluation results for a number of test subjects and analyze overall performance. Requiring only a smartphone or webcam, our method enables everyone to create their own fully animatable digital double, e.g., for social VR applications or virtual try-on for online fashion shopping. %K Computer Science, Computer Vision and Pattern Recognition, cs.CV
Bernard, F., Thunberg, J., Goncalves, J., and Theobalt, C. 2018. Synchronisation of Partial Multi-Matchings via Non-negative Factorisations. http://arxiv.org/abs/1803.06320.
(arXiv: 1803.06320)
Abstract
In this work we study permutation synchronisation for the challenging case of partial permutations, which plays an important role for the problem of matching multiple objects (e.g. images or shapes). The term synchronisation refers to the property that the set of pairwise matchings is cycle-consistent, i.e. in the full matching case all compositions of pairwise matchings over cycles must be equal to the identity. Motivated by clustering and matrix factorisation perspectives of cycle-consistency, we derive an algorithm to tackle the permutation synchronisation problem based on non-negative factorisations. In order to deal with the inherent non-convexity of the permutation synchronisation problem, we use an initialisation procedure based on a novel rotation scheme applied to the solution of the spectral relaxation. Moreover, this rotation scheme facilitates a convenient Euclidean projection to obtain a binary solution after solving our relaxed problem. In contrast to state-of-the-art methods, our approach is guaranteed to produce cycle-consistent results. We experimentally demonstrate the efficacy of our method and show that it achieves better results compared to existing methods.
Export
BibTeX
@online{Bernard_arXiv1803.06320, TITLE = {Synchronisation of Partial Multi-Matchings via Non-negative Factorisations}, AUTHOR = {Bernard, Florian and Thunberg, Johan and Goncalves, Jorge and Theobalt, Christian}, LANGUAGE = {eng}, URL = {http://arxiv.org/abs/1803.06320}, EPRINT = {1803.06320}, EPRINTTYPE = {arXiv}, YEAR = {2018}, ABSTRACT = {In this work we study permutation synchronisation for the challenging case of partial permutations, which plays an important role for the problem of matching multiple objects (e.g. images or shapes). The term synchronisation refers to the property that the set of pairwise matchings is cycle-consistent, i.e. in the full matching case all compositions of pairwise matchings over cycles must be equal to the identity. Motivated by clustering and matrix factorisation perspectives of cycle-consistency, we derive an algorithm to tackle the permutation synchronisation problem based on non-negative factorisations. In order to deal with the inherent non-convexity of the permutation synchronisation problem, we use an initialisation procedure based on a novel rotation scheme applied to the solution of the spectral relaxation. Moreover, this rotation scheme facilitates a convenient Euclidean projection to obtain a binary solution after solving our relaxed problem. In contrast to state-of-the-art methods, our approach is guaranteed to produce cycle-consistent results. We experimentally demonstrate the efficacy of our method and show that it achieves better results compared to existing methods.}, }
Endnote
%0 Report %A Bernard, Florian %A Thunberg, Johan %A Goncalves, Jorge %A Theobalt, Christian %+ External Organizations External Organizations External Organizations Computer Graphics, MPI for Informatics, Max Planck Society %T Synchronisation of Partial Multi-Matchings via Non-negative Factorisations : %G eng %U http://hdl.handle.net/21.11116/0000-0001-40BC-3 %U http://arxiv.org/abs/1803.06320 %D 2018 %X In this work we study permutation synchronisation for the challenging case of partial permutations, which plays an important role for the problem of matching multiple objects (e.g. images or shapes). The term synchronisation refers to the property that the set of pairwise matchings is cycle-consistent, i.e. in the full matching case all compositions of pairwise matchings over cycles must be equal to the identity. Motivated by clustering and matrix factorisation perspectives of cycle-consistency, we derive an algorithm to tackle the permutation synchronisation problem based on non-negative factorisations. In order to deal with the inherent non-convexity of the permutation synchronisation problem, we use an initialisation procedure based on a novel rotation scheme applied to the solution of the spectral relaxation. Moreover, this rotation scheme facilitates a convenient Euclidean projection to obtain a binary solution after solving our relaxed problem. In contrast to state-of-the-art methods, our approach is guaranteed to produce cycle-consistent results. We experimentally demonstrate the efficacy of our method and show that it achieves better results compared to existing methods. %K Computer Science, Computer Vision and Pattern Recognition, cs.CV,Mathematics, Optimization and Control, math.OC,Statistics, Machine Learning, stat.ML
Meka, A., Maximov, M., Zollhöfer, M., et al. 2018. LIME: Live Intrinsic Material Estimation. http://arxiv.org/abs/1801.01075.
(arXiv: 1801.01075)
Abstract
We present the first end to end approach for real time material estimation for general object shapes with uniform material that only requires a single color image as input. In addition to Lambertian surface properties, our approach fully automatically computes the specular albedo, material shininess, and a foreground segmentation. We tackle this challenging and ill posed inverse rendering problem using recent advances in image to image translation techniques based on deep convolutional encoder decoder architectures. The underlying core representations of our approach are specular shading, diffuse shading and mirror images, which allow to learn the effective and accurate separation of diffuse and specular albedo. In addition, we propose a novel highly efficient perceptual rendering loss that mimics real world image formation and obtains intermediate results even during run time. The estimation of material parameters at real time frame rates enables exciting mixed reality applications, such as seamless illumination consistent integration of virtual objects into real world scenes, and virtual material cloning. We demonstrate our approach in a live setup, compare it to the state of the art, and demonstrate its effectiveness through quantitative and qualitative evaluation.
Export
BibTeX
@online{Meka_arXiv1801.01075, TITLE = {LIME: {L}ive Intrinsic Material Estimation}, AUTHOR = {Meka, Abhimitra and Maximov, Maxim and Zollh{\"o}fer, Michael and Chatterjee, Avishek and Seidel, Hans-Peter and Richardt, Christian and Theobalt, Christian}, URL = {http://arxiv.org/abs/1801.01075}, EPRINT = {1801.01075}, EPRINTTYPE = {arXiv}, YEAR = {2018}, ABSTRACT = {We present the first end to end approach for real time material estimation for general object shapes with uniform material that only requires a single color image as input. In addition to Lambertian surface properties, our approach fully automatically computes the specular albedo, material shininess, and a foreground segmentation. We tackle this challenging and ill posed inverse rendering problem using recent advances in image to image translation techniques based on deep convolutional encoder decoder architectures. The underlying core representations of our approach are specular shading, diffuse shading and mirror images, which allow to learn the effective and accurate separation of diffuse and specular albedo. In addition, we propose a novel highly efficient perceptual rendering loss that mimics real world image formation and obtains intermediate results even during run time. The estimation of material parameters at real time frame rates enables exciting mixed reality applications, such as seamless illumination consistent integration of virtual objects into real world scenes, and virtual material cloning. We demonstrate our approach in a live setup, compare it to the state of the art, and demonstrate its effectiveness through quantitative and qualitative evaluation.}, }
Endnote
%0 Report %A Meka, Abhimitra %A Maximov, Maxim %A Zollhöfer, Michael %A Chatterjee, Avishek %A Seidel, Hans-Peter %A Richardt, Christian %A Theobalt, Christian %+ Computer Graphics, MPI for Informatics, Max Planck Society D2 External Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society External Organizations Computer Graphics, MPI for Informatics, Max Planck Society %T LIME: Live Intrinsic Material Estimation : %U http://hdl.handle.net/21.11116/0000-0001-40D9-2 %U http://arxiv.org/abs/1801.01075 %D 2018 %X We present the first end to end approach for real time material estimation for general object shapes with uniform material that only requires a single color image as input. In addition to Lambertian surface properties, our approach fully automatically computes the specular albedo, material shininess, and a foreground segmentation. We tackle this challenging and ill posed inverse rendering problem using recent advances in image to image translation techniques based on deep convolutional encoder decoder architectures. The underlying core representations of our approach are specular shading, diffuse shading and mirror images, which allow to learn the effective and accurate separation of diffuse and specular albedo. In addition, we propose a novel highly efficient perceptual rendering loss that mimics real world image formation and obtains intermediate results even during run time. The estimation of material parameters at real time frame rates enables exciting mixed reality applications, such as seamless illumination consistent integration of virtual objects into real world scenes, and virtual material cloning. We demonstrate our approach in a live setup, compare it to the state of the art, and demonstrate its effectiveness through quantitative and qualitative evaluation. %K Computer Science, Computer Vision and Pattern Recognition, cs.CV
Sun, Q., Tewari, A., Xu, W., Fritz, M., Theobalt, C., and Schiele, B. 2018. A Hybrid Model for Identity Obfuscation by Face Replacement. http://arxiv.org/abs/1804.04779.
(arXiv: 1804.04779)
Abstract
As more and more personal photos are shared and tagged in social media, avoiding privacy risks such as unintended recognition becomes increasingly challenging. We propose a new hybrid approach to obfuscate identities in photos by head replacement. Our approach combines state of the art parametric face synthesis with latest advances in Generative Adversarial Networks (GAN) for data-driven image synthesis. On the one hand, the parametric part of our method gives us control over the facial parameters and allows for explicit manipulation of the identity. On the other hand, the data-driven aspects allow for adding fine details and overall realism as well as seamless blending into the scene context. In our experiments, we show highly realistic output of our system that improves over the previous state of the art in obfuscation rate while preserving a higher similarity to the original image content.
Export
BibTeX
@online{Sun_arXiv1804.04779, TITLE = {A Hybrid Model for Identity Obfuscation by Face Replacement}, AUTHOR = {Sun, Qianru and Tewari, Ayush and Xu, Weipeng and Fritz, Mario and Theobalt, Christian and Schiele, Bernt}, LANGUAGE = {eng}, URL = {http://arxiv.org/abs/1804.04779}, EPRINT = {1804.04779}, EPRINTTYPE = {arXiv}, YEAR = {2018}, ABSTRACT = {As more and more personal photos are shared and tagged in social media, avoiding privacy risks such as unintended recognition becomes increasingly challenging. We propose a new hybrid approach to obfuscate identities in photos by head replacement. Our approach combines state of the art parametric face synthesis with latest advances in Generative Adversarial Networks (GAN) for data-driven image synthesis. On the one hand, the parametric part of our method gives us control over the facial parameters and allows for explicit manipulation of the identity. On the other hand, the data-driven aspects allow for adding fine details and overall realism as well as seamless blending into the scene context. In our experiments, we show highly realistic output of our system that improves over the previous state of the art in obfuscation rate while preserving a higher similarity to the original image content.}, }
Endnote
%0 Report %A Sun, Qianru %A Tewari, Ayush %A Xu, Weipeng %A Fritz, Mario %A Theobalt, Christian %A Schiele, Bernt %+ Computer Vision and Multimodal Computing, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Vision and Multimodal Computing, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Vision and Multimodal Computing, MPI for Informatics, Max Planck Society %T A Hybrid Model for Identity Obfuscation by Face Replacement : %G eng %U http://hdl.handle.net/21.11116/0000-0001-3F9B-B %U http://arxiv.org/abs/1804.04779 %D 2018 %X As more and more personal photos are shared and tagged in social media, avoiding privacy risks such as unintended recognition becomes increasingly challenging. We propose a new hybrid approach to obfuscate identities in photos by head replacement. Our approach combines state of the art parametric face synthesis with latest advances in Generative Adversarial Networks (GAN) for data-driven image synthesis. On the one hand, the parametric part of our method gives us control over the facial parameters and allows for explicit manipulation of the identity. On the other hand, the data-driven aspects allow for adding fine details and overall realism as well as seamless blending into the scene context. In our experiments, we show highly realistic output of our system that improves over the previous state of the art in obfuscation rate while preserving a higher similarity to the original image content. %K Computer Science, Computer Vision and Pattern Recognition, cs.CV,Computer Science, Cryptography and Security, cs.CR
Xu, W., Chatterjee, A., Zollhöfer, M., et al. 2018b. Mo2Cap2: Real-time Mobile 3D Motion Capture with a Cap-mounted Fisheye Camera. http://arxiv.org/abs/1803.05959.
(arXiv: 1803.05959)
Abstract
We propose the first real-time approach for the egocentric estimation of 3D human body pose in a wide range of unconstrained everyday activities. This setting has a unique set of challenges, such as mobility of the hardware setup, and robustness to long capture sessions with fast recovery from tracking failures. We tackle these challenges based on a novel lightweight setup that converts a standard baseball cap to a device for high-quality pose estimation based on a single cap-mounted fisheye camera. From the captured egocentric live stream, our CNN based 3D pose estimation approach runs at 60Hz on a consumer-level GPU. In addition to the novel hardware setup, our other main contributions are: 1) a large ground truth training corpus of top-down fisheye images and 2) a novel disentangled 3D pose estimation approach that takes the unique properties of the egocentric viewpoint into account. As shown by our evaluation, we achieve lower 3D joint error as well as better 2D overlay than the existing baselines.
Export
BibTeX
@online{Xu_arXiv1803.05959, TITLE = {{Mo2Cap2}: Real-time Mobile {3D} Motion Capture with a Cap-mounted Fisheye Camera}, AUTHOR = {Xu, Weipeng and Chatterjee, Avishek and Zollh{\"o}fer, Michael and Rhodin, Helge and Fua, Pascal and Seidel, Hans-Peter and Theobalt, Christian}, LANGUAGE = {eng}, URL = {http://arxiv.org/abs/1803.05959}, EPRINT = {1803.05959}, EPRINTTYPE = {arXiv}, YEAR = {2018}, ABSTRACT = {We propose the first real-time approach for the egocentric estimation of 3D human body pose in a wide range of unconstrained everyday activities. This setting has a unique set of challenges, such as mobility of the hardware setup, and robustness to long capture sessions with fast recovery from tracking failures. We tackle these challenges based on a novel lightweight setup that converts a standard baseball cap to a device for high-quality pose estimation based on a single cap-mounted fisheye camera. From the captured egocentric live stream, our CNN based 3D pose estimation approach runs at 60Hz on a consumer-level GPU. In addition to the novel hardware setup, our other main contributions are: 1) a large ground truth training corpus of top-down fisheye images and 2) a novel disentangled 3D pose estimation approach that takes the unique properties of the egocentric viewpoint into account. As shown by our evaluation, we achieve lower 3D joint error as well as better 2D overlay than the existing baselines.}, }
Endnote
%0 Report %A Xu, Weipeng %A Chatterjee, Avishek %A Zollhöfer, Michael %A Rhodin, Helge %A Fua, Pascal %A Seidel, Hans-Peter %A Theobalt, Christian %+ Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society External Organizations External Organizations Computer Graphics, MPI for Informatics, Max Planck Society Computer Graphics, MPI for Informatics, Max Planck Society %T Mo2Cap2: Real-time Mobile 3D Motion Capture with a Cap-mounted Fisheye Camera : %G eng %U http://hdl.handle.net/21.11116/0000-0001-3C65-B %U http://arxiv.org/abs/1803.05959 %D 2018 %X We propose the first real-time approach for the egocentric estimation of 3D human body pose in a wide range of unconstrained everyday activities. This setting has a unique set of challenges, such as mobility of the hardware setup, and robustness to long capture sessions with fast recovery from tracking failures. We tackle these challenges based on a novel lightweight setup that converts a standard baseball cap to a device for high-quality pose estimation based on a single cap-mounted fisheye camera. From the captured egocentric live stream, our CNN based 3D pose estimation approach runs at 60Hz on a consumer-level GPU. In addition to the novel hardware setup, our other main contributions are: 1) a large ground truth training corpus of top-down fisheye images and 2) a novel disentangled 3D pose estimation approach that takes the unique properties of the egocentric viewpoint into account. As shown by our evaluation, we achieve lower 3D joint error as well as better 2D overlay than the existing baselines. %K Computer Science, Computer Vision and Pattern Recognition, cs.CV