Continuous Trade-off Optimization between Fast and Accurate Deep Face Detectors
This work addresses the efficiency problem for face detection applications, but it is incremental as it builds on existing detectors with straightforward splitting methods.
The paper tackled the trade-off between speed and accuracy in deep face detectors by proposing five approaches that split test images into easy and difficult batches, using faster and slower detectors respectively, and showed that these difficulty metrics outperform random splits on AFW and FDDB datasets.
Although deep neural networks offer better face detection results than shallow or handcrafted models, their complex architectures come with higher computational requirements and slower inference speeds than shallow neural networks. In this context, we study five straightforward approaches to achieve an optimal trade-off between accuracy and speed in face detection. All the approaches are based on separating the test images in two batches, an easy batch that is fed to a faster face detector and a difficult batch that is fed to a more accurate yet slower detector. We conduct experiments on the AFW and the FDDB data sets, using MobileNet-SSD as the fast face detector and S3FD (Single Shot Scale-invariant Face Detector) as the accurate face detector, both models being pre-trained on the WIDER FACE data set. Our experiments show that the proposed difficulty metrics compare favorably to a random split of the images.