Few-shot object detection (FSOD) aims to detect objects using only few examples. It's critically needed for many practical applications but so far remains challenging. We propose a meta-learning based few-shot object detection method by transferring meta-knowledge learned from data-abundant base classes to data-scarce novel classes. Our method incorporates a coarse-to-fine approach into the proposal based object detection framework and integrates prototype based classifiers into both the proposal generation and classification stages. To improve proposal generation for few-shot novel classes, we propose to learn a lightweight matching network to measure the similarity between each spatial position in the query image feature map and spatially-pooled class features, instead of the traditional object/nonobject classifier, thus generating category-specific proposals and improving proposal recall for novel classes. To address the spatial misalignment between generated proposals and few-shot class examples, we propose a novel attentive feature alignment method, thus improving the performance of few-shot object detection. Meanwhile we jointly learn a Faster R-CNN detection head for base classes. Extensive experiments conducted on multiple FSOD benchmarks show our proposed approach achieves state of the art results under (incremental) few-shot learning settings.

中文翻译：

---

Meta Faster R-CNN：通过细心的特征对准实现精确的少量目标检测

很少有物体检测（FSOD）旨在仅使用几个示例来检测物体。许多实际应用都非常需要它，但到目前为止仍然充满挑战。通过将从数据丰富的基类中学到的元知识转移到数据稀缺的新颖类中，我们提出了一种基于元学习的少发对象检测方法。我们的方法在基于提议的对象检测框架中采用了从粗到精的方法，并将基于原型的分类器集成到提议的生成和分类阶段。为了改善少拍新颖类的提议生成，我们提议学习一种轻量级的匹配网络来测量查询图像特征图中每个空间位置与空间合并类特征之间的相似性，而不是传统的对象/非对象分类器，从而生成针对特定类别的提案，并改善新颖类的提案回忆。为了解决生成的建议和少拍类示例之间的空间不对准问题，我们提出了一种新颖的关注特征对准方法，从而提高了少拍目标检测的性能。同时，我们共同学习了针对基类的Faster R-CNN检测头。在多个FSOD基准上进行的大量实验表明，我们提出的方法在（渐进式）少拍学习设置下可以达到最新的结果。同时，我们共同学习了针对基类的Faster R-CNN检测头。在多个FSOD基准上进行的大量实验表明，我们提出的方法在（渐进式）少拍学习设置下可以达到最新的结果。同时，我们共同学习了针对基类的Faster R-CNN检测头。在多个FSOD基准上进行的大量实验表明，我们提出的方法在（渐进式）少拍学习设置下可以达到最新的结果。