FCIS方法的描述，实例分割其它相关描述
- 1.State-of the-art approaches to instance segmentation like Mask RCNN [18] and FCIS [24] directly build off of advances in object detection like Faster R-CNN [37] and R-FCN [8].来源于yolact
重点关注什么是R-FCN ，
R-FCN: Object detection via region-based fully convolutional networks.
2.mask rcnn中也提到了2. Related Work
Most recently, Li et al. [21] combined the segment pro
posal system in [5] and object detection system in [8] for
“fully convolutional instance segmentation” (FCIS). The
common idea in [5, 8, 21] is to predict a set of position
sensitive output channels fully convolutionally. These
channels simultaneously address object classes, boxes, and
masks, making the system fast. But FCIS exhibits system
atic errors on overlapping instances and creates spurious
edges (Figure 5), showing that it is challenged by the fun
damental diffificulties of segmenting instances.
先分割再聚类这种方法慢，segmentation precedes recognition, which is slow and less accurate.（出自mask rcnn）
mask rcnn是两阶段的也慢，出自yolact论文