Progressive structure network-based multiscale feature fusion for object detection in real-time application
Progressive structure network-based multiscale feature fusion for object detection in real-time application
复制标题
DOI：
10.1016/j.engappai.2021.104486
复制
发表时间：
2021-10-09
影响因子：
8
通讯作者：
Yu, Haibin
中科院分区：
文献类型：
--
作者：
Wang, Haifeng;Jiang, Lvjiyuan;Yu, Haibin
基于深度学习的目标检测技术已经对我们的日常生活产生了广泛的影响。目前，特征金字塔是一种广泛使用的多尺度目标检测技术，该技术的有效性已经得到证明。然而，金字塔结构中存在多尺度特征对齐、融合后模型混乱、特征冗余、非局部特征融合等问题。在本文中，我们提出了一种新颖的渐进结构网络来解决上述问题。所提出的结构包含三个模块：多尺度特征对齐融合、不同尺度通道和空间位置自适应加权融合以及多尺度全局和局部特征融合。所提出的结构能够更有效地融合来自不同特征层的信息。随后，可以减少不同尺度之间的语义差距。此外，所提出的结构可以保持检测网络的稳定性，并且通过与其他最先进的特征融合方法进行比较，其性能得到了证明。所提出的渐进式网络结构也已应用于实际的目标检测任务，并验证了我们方法的实际应用有效性。
Deep learning-based target detection techniques have already made a wide-range impact on our daily life. Currently, a feature pyramid is a widely utilized technique for multiscale target detection, the effectiveness of the technique has already been proved. Nevertheless, in the pyramid structure, problems, such as multiscale feature alignment, model turmoil after fusion, feature redundancy, and no-local feature fusion, exist. In this paper, we propose a novel progressive structure network to solve the aforementioned problems. The proposed structure contains three modules: multiscale feature alignment fusion, different scale channels & spatial location adaptive weighted fusion, and multiscale global and local feature fusion. The proposed structure is capable of fusing information from different feature layers more effectively. Subsequently, the semantic gaps among different scales can be reduced. Furthermore, the proposed structure can maintain the stability of the detection network and its performance has been proved by comparing with other state-of-art feature fusion method. The proposed progressive network structure has also been applied to actual target detection tasks and the practical application effectiveness of our method has been verified.