Supervision-by-Registration: An Unsupervised Approach to Improve the Precision of Facial Landmark Detectors

@article{Dong2018SupervisionbyRegistrationAU,
  title={Supervision-by-Registration: An Unsupervised Approach to Improve the Precision of Facial Landmark Detectors},
  author={Xuanyi Dong and Shoou-I Yu and Xinshuo Weng and Shih-En Wei and Yi Yang and Yaser Sheikh},
  journal={2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2018},
  pages={360-368}
}
  • Xuanyi Dong, Shoou-I Yu, +3 authors Yaser Sheikh
  • Published 2018
  • Computer Science
  • 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition
  • In this paper, we present supervision-by-registration, an unsupervised approach to improve the precision of facial landmark detectors on both images and video. Our key observation is that the detections of the same landmark in adjacent frames should be coherent with registration, i.e., optical flow. Interestingly, coherency of optical flow is a source of supervision that does not require manual labeling, and can be leveraged during detector training. For example, we can enforce in the training… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    Citations

    Publications citing this paper.
    SHOWING 1-10 OF 88 CITATIONS, ESTIMATED 89% COVERAGE

    Multiview Supervision By Registration

    VIEW 8 EXCERPTS
    CITES BACKGROUND, METHODS & RESULTS
    HIGHLY INFLUENCED

    Towards Omni-Supervised Face Alignment for Large Scale Unlabeled Videos

    VIEW 15 EXCERPTS
    CITES METHODS, BACKGROUND & RESULTS
    HIGHLY INFLUENCED

    FAB: A Robust Facial Landmark Detection Framework for Motion-Blurred Videos

    Robust Facial Landmark Detection via Occlusion-Adaptive Deep Networks

    VIEW 3 EXCERPTS
    CITES BACKGROUND & METHODS

    Deep, Landmark-Free FAME: Face Alignment, Modeling, and Expression Estimation

    VIEW 3 EXCERPTS
    CITES BACKGROUND & METHODS

    Laplace Landmark Localization

    VIEW 4 EXCERPTS
    CITES BACKGROUND & METHODS
    HIGHLY INFLUENCED

    FILTER CITATIONS BY YEAR

    2018
    2020

    CITATION STATISTICS

    • 11 Highly Influenced Citations

    • Averaged 29 Citations per year from 2018 through 2020

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 44 REFERENCES

    The First Facial Landmark Tracking in-the-Wild Challenge: Benchmark and Results

    Offline Deformable Face Tracking in Arbitrary Videos

    Face tracking and recognition with visual constraints in real-world videos

    Annotated Facial Landmarks in the Wild: A large-scale, real-world database for facial landmark localization

    VIEW 7 EXCERPTS
    HIGHLY INFLUENTIAL

    300 Faces in-the-Wild Challenge: The First Facial Landmark Localization Challenge

    VIEW 3 EXCERPTS

    A Deep Regression Architecture with Two-Stage Re-initialization for High Performance Facial Landmark Detection

    VIEW 6 EXCERPTS
    HIGHLY INFLUENTIAL

    Personalizing Human Video Pose Estimation

    VIEW 3 EXCERPTS

    Supervised Descent Method and Its Applications to Face Alignment

    VIEW 3 EXCERPTS