DCNet (CVPR. 2021)

1. Motivation

这篇文章中，作者针对图1出现的2个问题，指出，由于support 和query img 之间的关系没法完全提取，因为之前的方法都是使用GAP的方法，没有考虑局部信息。其次，对于分类估计错误，以及遮挡的问题，

• Firstly, relations between support fea- tures and query feature are hardly fully explored in previous few-shot detection works, where global pooling opera- tion on support features is mostly adopted to modulate the query branch, which is prone to loss of detailed local context.
• Specifically, appearance changes and occlusions are common for objects, as shown Fig. 1.

2. Contribution

• We propose a dense relation distillation module for few-shot detection problem, which targets at fully ex- ploiting support information to assist the detection pro- cess for objects from novel classes.
• We propose an adaptive context-aware feature aggre- gation module to better capture global and local fea- tures to alleviate scale variation problem, boosting the performance of few-shot detection.
• Extensive experiments illustrate that our approach has achieved a consistent improvement on PASCAL VOC and MS COCO datasets. Specially, our approach achieves better performance than the state-of-the-art methods on the two datasets.

3. Method

4. Experiments

4.1 VOC

4.2 COCO

4.3 Impact of context-aware feature aggregation module

4.4 Impact of different RoI pooling resolutions.