BlazeFace: Sub-millisecond Neural Face Detection on Mobile GPUs

Abstract

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 a lightweight feature extraction network inspired by, but distinct from MobileNetV1/V2, a GPU-friendly anchor scheme modified from Single Shot MultiBox Detector (SSD), and an improved tie resolution strategy alternative to non-maximum suppression.

4 pages, 3 figures; CVPR Workshop on Computer Vision for Augmented and Virtual Reality, Long Beach, CA, USA, 2019

References in corpus (3)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• Pooling Pyramid Network for Object Detection

Cited by in corpus (18)

• MediaPipe Hands: On-device Real-time Hand Tracking
• Deepfake Video Detection Using Convolutional Vision Transformer
• BlazePose: On-device Real-time Body Pose tracking
• Fast Facial Landmark Detection and Applications: A Survey
• Going Deeper Into Face Detection: A Survey
• BakedAvatar: Baking Neural Fields for Real-Time Head Avatar Synthesis
• Real-time Pupil Tracking from Monocular Video for Digital Puppetry
• Investigating the Impact of Pre-processing and Prediction Aggregation on the DeepFake Detection Task
• Edge Devices Inference Performance Comparison
• A multimodal dataset for understanding the impact of mobile phones on remote online virtual education
• Smart at what cost? Characterising Mobile Deep Neural Networks in the wild
• Efficient Memory Management for Deep Neural Net Inference
• Neonatal Face and Facial Landmark Detection from Video Recordings
• Evaluating User Experience and Data Quality in Gamified Data Collection for Appearance-Based Gaze Estimation
• What's wrong with this video? Comparing Explainers for Deepfake Detection
• Skin-MIMO: Vibration-based MIMO Communication over Human Skin
• Towards Solving the DeepFake Problem : An Analysis on Improving DeepFake Detection using Dynamic Face Augmentation
• Perceptual Gradient Networks