DetectoRS: Detecting Objects with Recursive Feature Pyramid and Switchable Atrous Convolution
Siyuan Qiao1 Liang-Chieh Chen2 Alan Yuille1 1Johns Hopkins University 2Google Research
Abstract
Many modern object detectors demonstrate outstanding performances by using the mechanism of looking and thinking twice. In this paper, we explore this mechanism in the backbone design for object detection. At the macro level, we propose Recursive Feature Pyramid, which incorporates extra feedback connections from Feature Pyramid Networks into the bottom-up backbone layers. At the micro level, we propose Switchable Atrous Convolution, which convolves the features with different atrous rates and gathers the results using switch functions. Combining them results in DetectoRS, which significantly improves the performances of object detection. On COCO test-dev, DetectoRS achieves state-of-the-art 55.7% box AP for object detection, 48.5% mask AP for instance segmentation, and 50.0% PQ for panoptic segmentation. The code is made publicly available111 https://github.com/joe-siyuan-qiao/DetectoRS.
中文速览
检测人眼识别物体时会"多看几眼再做判断"，但这一机制此前很少被系统地引入检测器的骨干网络设计中。作者提出了 DetectoRS，在宏观层面设计了递归特征金字塔（Recursive Feature Pyramid, RFP），让特征金字塔网络的高层语义通过反馈连接重新流回底层骨干，使整个网络对图像"看两遍甚至更多遍"，从而生成越来越强的多尺度表示；在微观层面设计了可切换空洞卷积（Switchable Atrous Convolution, SAC），用空间自适应的开关函数动态融合不同膨胀率下的卷积结果，且可以直接从预训练权重平滑迁移而无需从头训练。将两者结合后，DetectoRS 在 COCO 测试集上分别达到 55.7% 框 AP、48.5% 掩码 AP 和 50.0% 全景分割 PQ，全面刷新了当时的最优记录，证明"看两遍"的思想在骨干网络层面同样能带来显著收益。
原文 arXiv:2006.02334;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2006.02334v2