An anchor-free region proposal network for Faster R-CNN-based text detection approaches

The anchor mechanism of Faster R-CNN and SSD framework is considered not effective enough to scene text detection, which can be attributed to its Intersection-over-Union-based matching criterion between anchors and ground-truth boxes. In order to better enclose scene text instances of various shapes, it requires to design anchors of various scales, aspect ratios and even orientations manually, which makes anchor-based methods sophisticated and inefficient. In this paper, we propose a novel anchor-free region proposal network (AF-RPN) to replace the original anchor-based RPN in the Faster R-CNN framework to address the above problem. Compared with the anchor-based region proposal generation approaches (e.g., RPN, FPN-RPN, RRPN and FPN-RRPN), AF-RPN can get rid of complicated anchor design and achieves higher recall rate on both horizontal and multi-oriented text detection benchmark tasks. Owing to the high-quality text proposals, our Faster R-CNN-based two-stage text detection approach achieves the state-of-the-art results on ICDAR-2017 MLT, COCO-Text, ICDAR-2015 and ICDAR-2013 text detection benchmark tasks by only using single-scale and single-model testing.