FA-RPN: Floating Region Proposals for Face Detection
This work addresses efficiency and accuracy improvements in face detection, an incremental advancement for computer vision applications.
The paper tackles the computational bottleneck of generating region proposals for face detection by proposing an efficient anchor placement strategy and a pooling-based approach, resulting in a face detector that achieves 89.4% mAP on the WIDER dataset.
We propose a novel approach for generating region proposals for performing face-detection. Instead of classifying anchor boxes using features from a pixel in the convolutional feature map, we adopt a pooling-based approach for generating region proposals. However, pooling hundreds of thousands of anchors which are evaluated for generating proposals becomes a computational bottleneck during inference. To this end, an efficient anchor placement strategy for reducing the number of anchor-boxes is proposed. We then show that proposals generated by our network (Floating Anchor Region Proposal Network, FA-RPN) are better than RPN for generating region proposals for face detection. We discuss several beneficial features of FA-RPN proposals like iterative refinement, placement of fractional anchors and changing anchors which can be enabled without making any changes to the trained model. Our face detector based on FA-RPN obtains 89.4% mAP with a ResNet-50 backbone on the WIDER dataset.