Scale-Aware Face Detection
This work addresses the computational bottleneck in face detection for applications requiring real-time or resource-efficient processing, representing an incremental improvement over existing methods.
The paper tackles the inefficiency of CNN-based face detectors in handling diverse face scales by proposing a Scale-aware Face Detector (SAFD) that predicts scale distributions to guide image zooming, achieving better performance with less computational cost, as shown in experiments on FDDB, MALF, and AFW datasets.
Convolutional neural network (CNN) based face detectors are inefficient in handling faces of diverse scales. They rely on either fitting a large single model to faces across a large scale range or multi-scale testing. Both are computationally expensive. We propose Scale-aware Face Detector (SAFD) to handle scale explicitly using CNN, and achieve better performance with less computation cost. Prior to detection, an efficient CNN predicts the scale distribution histogram of the faces. Then the scale histogram guides the zoom-in and zoom-out of the image. Since the faces will be approximately in uniform scale after zoom, they can be detected accurately even with much smaller CNN. Actually, more than 99% of the faces in AFW can be covered with less than two zooms per image. Extensive experiments on FDDB, MALF and AFW show advantages of SAFD.