Knowledge Commons of Institute of Automation,CAS
AugFPN: Improving Multi-scale Feature Learning for Object Detection | |
Guo, Chaoxu1,2; Fan, Bin1; Zhang, Qian3; Xiang, Shiming1,2; Pan, Chunhong1 | |
2020-06 | |
会议名称 | IEEE Proceedings of Computer Vision and Pattern Recognition |
页码 | 12595-12604 |
会议日期 | 2020-06-14 |
会议地点 | online meeting |
摘要 | Current state-of-the-art detectors typically exploit feature pyramid to detect objects at different scales. Among them, FPN is one of the representative works that build a feature pyramid by multi-scale features summation. However, the design defects behind prevent the multi-scale features from being fully exploited. In this paper, we begin by first analyzing the design defects of feature pyramid in FPN, and then introduce a new feature pyramid architecture named AugFPN to address these problems. Specifically, AugFPN consists of three components: Consistent Supervision, Residual Feature Augmentation, and Soft RoI Selection. AugFPN narrows the semantic gaps between features of different scales before feature fusion through Consistent Supervision. In feature fusion, ratio-invariant context information is extracted by Residual Feature Augmentation to reduce the information loss of feature map at the highest pyramid level. Finally, Soft RoI Selection is employed to learn a better RoI feature adaptively after feature fusion. By replacing FPN with AugFPN in Faster RCNN, our models achieve 2.3 and 1.6 points higher Average Precision (AP) when using ResNet50 and MobileNet-v2 as backbone respectively. Furthermore, AugFPN improves RetinaNet by 1.6 points AP and FCOS by 0.9 points AP when using ResNet50 as backbone. |
关键词 | AugFPN, Object Detection |
学科门类 | 工学 |
收录类别 | EI |
资助项目 | National Science Foundation of China[61573352,61876180] ; Major Project for New Generation of AI[2018AAA0100400] ; Young Elite Scientists Sponsorship Program by CAST[2018QNRC001] ; Beijing Natural Science Foundation[4162064] ; National Natural Science Foundation of China[91646207] ; National Natural Science Foundation of China[61773377] ; National Natural Science Foundation of China[61773377] ; National Natural Science Foundation of China[91646207] ; Beijing Natural Science Foundation[4162064] ; Young Elite Scientists Sponsorship Program by CAST[2018QNRC001] ; Major Project for New Generation of AI[2018AAA0100400] ; National Science Foundation of China[61573352,61876180] |
语种 | 英语 |
七大方向——子方向分类 | 目标检测、跟踪与识别 |
文献类型 | 会议论文 |
条目标识符 | http://ir.ia.ac.cn/handle/173211/39183 |
专题 | 多模态人工智能系统全国重点实验室_先进时空数据分析与学习 |
通讯作者 | Fan, Bin |
作者单位 | 1.Institute of Automation, Chinese Academy of Science 2.School of Artifical Intelligence, University of Chinese Academy of Science 3.Horizon Robotics |
推荐引用方式 GB/T 7714 | Guo, Chaoxu,Fan, Bin,Zhang, Qian,et al. AugFPN: Improving Multi-scale Feature Learning for Object Detection[C],2020:12595-12604. |
条目包含的文件 | 下载所有文件 | |||||
文件名称/大小 | 文献类型 | 版本类型 | 开放类型 | 使用许可 | ||
Guo_AugFPN_Improving(631KB) | 会议论文 | 开放获取 | CC BY-NC-SA | 浏览 下载 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论