SAFNet: A Semi-Anchor-Free Network with Enhanced Feature Pyramid for Object Detection.
Zhenchao Jin,Bin Liu,Qi Chu,Nenghai Yu
DOI: https://doi.org/10.1109/tip.2020.3028196
IF: 10.6
2020-01-01
IEEE Transactions on Image Processing
Abstract:In recent years, the field of object detection has made significant progress. The success of most state-of-the-art object detectors is derived from the use of feature pyramid and the carefully designed anchor boxes. However, the existing methods for constructing feature pyramid blindly integrate multi-scale representations on each feature hierarchy. Furthermore, these detectors also suffer from some drawbacks brought by the hand-designed anchors. To mitigate the adverse effects caused thereby, we propose a semi-anchor-free network with enhanced feature pyramid for object detection, named SAFNet. Specifically, to better construct feature pyramid, we propose a novel enhanced feature pyramid generation paradigm, which consists of two modules, i.e., adaptive feature fusion module (AFFM) and self-enhanced module (SEM). The paradigm adaptively integrates multi-scale representations in a non-linear way meanwhile suppresses the redundant semantic information for each pyramid level. Thus, a clean and enhanced feature pyramid could be obtained. In addition, an adaptive anchor generator (AAG) is designed to yield fewer but more suitable anchor boxes for each input image. Benefiting from the enhanced feature pyramid, AAG is capable of generating more accurate anchor boxes by introducing few priors. With this semi-anchor-free method, our detector has the ability to alleviate the drawbacks of hand-designed anchors meanwhile retain the merits of anchor-based methods. Extensive experiments demonstrate the effectiveness of our approach. Profited from the proposed modules, SAFNet significantly boosts the detection performance, i.e., achieving 2 points and 2.1 points higher Average Precision (AP) than RetinaNet on PASCAL VOC and MS COCO respectively.