Sample and Computation Redistribution for Efficient Face Detection

Abstract

Although tremendous strides have been made in uncontrolled face detection, efficient face detection with a low computation cost as well as high precision remains an open challenge. In this paper, we point out that training data sampling and computation distribution strategies are the keys to efficient and accurate face detection. Motivated by these observations, we introduce two simple but effective methods (1) Sample Redistribution (SR), which augments training samples for the most needed stages, based on the statistics of benchmark datasets; and (2) Computation Redistribution (CR), which reallocates the computation between the backbone, neck and head of the model, based on a meticulously defined search methodology. Extensive experiments conducted on WIDER FACE demonstrate the state-of-the-art efficiency-accuracy trade-off for the proposed \scrfd family across a wide range of compute regimes. In particular, \scrfdf{34} outperforms the best competitor, TinaFace, by $3.86\%$ (AP at hard set) while being more than \emph{3$\times$ faster} on GPUs with VGA-resolution images. We also release our code to facilitate future research.

References in corpus (6)

• MMDetection: Open MMLab Detection Toolbox and Benchmark
• RetinaFace: Single-stage Dense Face Localisation in the Wild
• TinaFace: Strong but Simple Baseline for Face Detection
• Computation Reallocation for Object Detection
• HAMBox: Delving into Online High-quality Anchors Mining for Detecting Outer Faces
• ASFD: Automatic and Scalable Face Detector

Cited by in corpus (6)

• YOLO5Face: Why Reinventing a Face Detector
• Whose Emotion Matters? Speaking Activity Localisation without Prior Knowledge
• End-to-end Evaluation of Practical Video Analytics Systems for Face Detection and Recognition
• Double Trouble? Impact and Detection of Duplicates in Face Image Datasets
• Embedding Aggregation for Forensic Facial Comparison
• Eye Sclera for Fair Face Image Quality Assessment