arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1707.08682cs.CV
收藏

上下文感知单次检测器

Context-Aware Single-Shot Detector

Futurewei Technologies(未来wei技术公司);University of Texas at Arlington(德克萨斯大学阿灵顿分校)

Wei Xiang, Dong-Qing Zhang, Heather Yu, Vassilis Athitsos

摘要:

SSD是最先进的目标检测算法之一,它结合了高检测精度和实时速度。然而,人们普遍认为,SSD在检测小对象方面不如大对象准确,因为它忽略了提案框之外的上下文。在本文中,我们提出了CSSD-一个速记上下文感知的单镜头多框对象检测器。CSSD是建立在SSD之上的,额外的层建模多尺度上下文。我们描述了CSSD的两种变体,它们在上下文层上不同,分别使用扩张卷积层(DiCSSD)和去卷积层(DeCSSD)。实验结果表明,多尺度上下文建模显著提高了检测精度。此外,我们研究了有效感受野(ERFs)和理论感受野(TRFs)之间的关系,特别是在VGGNet上。实验结果进一步加强了我们的结论,SSD与上下文层结合实现了更好的检测结果,特别是对于小对象(与最新的SSD相比,MS-COCO上的$+3.2\% {\rm AP}_{@0.5}$),同时保持了相当的运行时性能。

英文摘要:

SSD is one of the state-of-the-art object detection algorithms, and it combines high detection accuracy with real-time speed. However, it is widely recognized that SSD is less accurate in detecting small objects compared to large objects, because it ignores the context from outside the proposal boxes. In this paper, we present CSSD--a shorthand for context-aware single-shot multibox object detector. CSSD is built on top of SSD, with additional layers modeling multi-scale contexts. We describe two variants of CSSD, which differ in their context layers, using dilated convolution layers (DiCSSD) and deconvolution layers (DeCSSD) respectively. The experimental results show that the multi-scale context modeling significantly improves the detection accuracy. In addition, we study the relationship between effective receptive fields (ERFs) and the theoretical receptive fields (TRFs), particularly on a VGGNet. The empirical results further strengthen our conclusion that SSD coupled with context layers achieves better detection results especially for small objects ($+3.2\% {\rm AP}_{@0.5}$ on MS-COCO compared to the newest SSD), while maintaining comparable runtime performance.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑