In this thesis, we tackle the problem of designing a multi-view facial landmark detector which is robust and works in real-time on low-end hardware. Our landmark detector is an instance of the structured output classi ers describing the face by a mixture of tree based Deformable Part Models (DPM). We propose to learn parameters of the detector by the Structured Output Support Vector Machine algorithm which, in contrast to existing methods, directly optimizes a loss function closely related to the standard evaluation metrics used in landmark detection. We also propose a novel two-stage approach to learn the multi-view landmark detectors, which provides better localization accuracy and signi cantly reduces the overall learning time. We propose several speedups that enable to use the globally optimal prediction strategy based on the dynamic programming in real time even for dense landmark sets. The empirical evaluation shows that the proposed detector is competitive with the current state-ofthe-art both regarding the accuracy and speed.

We also propose two improvements of the Bundle Method for Regularized Risk Minimization (BMRM) algorithm which is among the most popular batch solvers used in structured output learning. First, we propose to augment the objective function by a quadratic prox-center whose strength is controlled by a novel adaptive strategy preventing zig-zag behavior in the cases when the genuine regularization term is weak. Second, we propose to speed up convergence by using multiple cutting plane models which better approximate the objective function with minimal increase in the computational cost. Experimental evaluation shows that the new BMRM algorithm which uses both improvements speeds up learning up to an order of magnitude on standard computer vision benchmarks, and 3 to 4 times when applied to the learning of the DPM based landmark detector.

Publications

  • M. Uricar, P. Krizek, D. Hurych, I. Sobh, S. Yogamani, P. Denny. Yes, we GAN: Applying Adversarial Techniques for Autohomous Driving. IS&T International Symposium on Electronic Imaging, Autonomous Vehicles and Machines 2019, 2019.
  • M. Uricar, P. Krizek, D. Hurych, S. Yogamani. Towards Optimal Design of Datasets and Validation Scheme for Autonomous Driving. CRACT: Critiquing and Correcting Trends in Machine Learning; Thirty-second Conference on Neural Information Processing Systems (NIPS), Workshops, (NIPSw) 2018.
    pdf
  • M. Uricar, R. Timofte, R. Rothe, J. Matas. Structured Output SVM Prediction of Apparent Age, Gender and Smile From Deep Features. The 29th IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Workshops, (CVPRw) 2016
    Received 3rd place in ChaLearn LAP Challenge, Track 1: Age Estimation
    pdf
  • M. Uricar, V. Franc, D. Thomas, A. Sugimoto, and V. Hlavac. Multi-view facial landmark detector learned by the Structured Output SVM. Image and Vision Computing, 2016.
    ScienceDirect link
  • J. Cech, V. Franc, M. Uricar, and J. Matas. Multi-view facial landmark detection by using a 3D shape model. Image and Vision Computing, 2016.
    ScienceDirect link
  • M. Uricar, V. Franc, and V. Hlavac. Facial Landmark Tracking by Tree-Based Deformable Part Model Based Detector. 300VW: The IEEE International Conference on Computer Vision (ICCV) Workshops, (ICCVw) 2015.
    pdf
  • M. Uricar, V. Franc, D. Thomas, A. Sugimoto, and V. Hlavac. Real-time Multi-view Facial Landmark Detector Learned by the Structured Output SVM. BWILD '15: International Workshop on Biometrics in the Wild, in conjunction with International Conference on Automatic Face and Gesture Recognition, (FGw) 2015
    pdf
  • M. Uricar, V. Franc, and V. Hlavac. Bundle Methods for Structured Output Learning - Back to the Roots, In Scandinavian Conference on Image Analysis, (SCIA), 2013.
    Springer Link
  • M. Uricar, V. Franc and V. Hlavac. Detector of Facial Landmarks Learned by the Structured Output SVM. VISAPP '12: Proceedings of the 7th International Conference on Computer Vision Theory and Applications, 2012.
    Received Best Paper Award
    pdf
  • M. Uricar and V. Franc. Efficient Algorithm for Regularized Risk Minimization. CVWW '12: Proceedings of the Computer Vision Winter Workshop, (CVWW), 2012.
    pdf

Download

Thesis

PhD Thesis - Multi-view Facial Landmark Detection

Defense Slides

Slides from the PhD defense

CV

Author's Curriculum Vitae

List of publications

Author's list of publications

Review 1

Review of the thesis by prof. Josef Kittler

Review 2

Review of the thesis by doc. Barbara Zitova

Review 3

Review of the thesis by prof. Radu Horaud

Get In Touch.

Are you interested in some details about the thesis? Or do you want to ask some questions? Feel free to use the following form to contact us.

Error boy
Your message was sent, thank you!