While recent progress has significantly boosted few-shot classification (FSC) performance, few-shot object detection (FSOD) remains challenging for modern learning systems. Existing FSOD systems follow FSC approaches, ignoring critical issues such as spatial variability and uncertain representations, and consequently result in low performance. Observing this, we propose a novel \textbf{Dual-Awareness Attention (DAnA)} mechanism that enables networks to adaptively interpret the given support images. DAnA transforms support images into \textbf{query-position-aware} (QPA) features, guiding detection networks precisely by assigning customized support information to each local region of the query. In addition, the proposed DAnA component is flexible and adaptable to multiple existing object detection frameworks. By adopting DAnA, conventional object detection networks, Faster R-CNN and RetinaNet, which are not designed explicitly for few-shot learning, reach state-of-the-art performance in FSOD tasks. In comparison with previous methods, our model significantly increases the performance by 47\% (+6.9 AP), showing remarkable ability under various evaluation settings.
翻译:虽然最近的进展大大提升了微小的分类(FSC)性能,但微小的物体探测(FSOD)对现代学习系统仍然具有挑战性。现有的FSOD系统采用FSC方法,忽略空间变异性和不确定表现等关键问题,从而导致低性能。我们注意到这一点,我们建议建立一个新型的\textbf{Dual-Awarnity 注意(DannA)机制,使网络能够对给定的支持图像进行适应性解释。DAnA将支持图像转换成\textf{query-position-aware}(QPA)功能,通过向每个查询的当地区域提供定制的支持信息来准确指导探测网络。此外,拟议的DAAA组件具有灵活性,适应多种现有的物体探测框架。通过采用DAAA、常规物体探测网络、更快的R-CNN和RetinaNet,这些网络并非专门设计用于微小的学习,在FSOD任务中达到最先进的性能。与以前的方法相比,我们的模型大大提高了47 ⁇ (+6. AP)的性能在各种评价环境中显示惊人的能力。