Corpus ID: 195886588

BlazeFace: Sub-millisecond Neural Face Detection on Mobile GPUs

@article{Bazarevsky2019BlazeFaceSN,
  title={BlazeFace: Sub-millisecond Neural Face Detection on Mobile GPUs},
  author={Valentin Bazarevsky and Yury Kartynnik and Andrey Vakunov and K. Raveendran and M. Grundmann},
  journal={ArXiv},
  year={2019},
  volume={abs/1907.05047}
}
  • Valentin Bazarevsky, Yury Kartynnik, +2 authors M. Grundmann
  • Published 2019
  • Computer Science
  • ArXiv
  • We present BlazeFace, a lightweight and well-performing face detector tailored for mobile GPU inference. It runs at a speed of 200-1000+ FPS on flagship devices. This super-realtime performance enables it to be applied to any augmented reality pipeline that requires an accurate facial region of interest as an input for task-specific models, such as 2D/3D facial keypoint or geometry estimation, facial features or expression classification, and face region segmentation. Our contributions include… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    The CpG dinucleotide and human genetic disease
    506
    Deepfakes Detection with Automatic Face Weighting
    2
    Efficient Memory Management for Deep Neural Net Inference
    Learning Illumination from Diverse Portraits
    MediaPipe Hands: On-device Real-time Hand Tracking

    References

    Publications referenced by this paper.
    SHOWING 1-7 OF 7 REFERENCES
    MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
    4728
    SSD: Single Shot MultiBox Detector
    7812
    MobileNetV2: Inverted Residuals and Linear Bottlenecks
    2303
    Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
    17487
    Pooling Pyramid Network for Object Detection
    5
    Fast R-CNN
    8461
    The Pascal Visual Object Classes (VOC) Challenge
    7927