Instance Segmentation Models

BlendMask is an instance segmentation framework built on top of the FCOS object detector. The bottom module uses either backbone or FPN features to predict a set of bases. A single convolution layer is added on top of the detection towers to produce attention masks along with each bounding box prediction. For each predicted instance, the blender crops the bases with its bounding box and linearly combine them according the learned attention maps. Note that the Bottom Module can take features either from ‘C’, or ‘P’ as the input.

Source: BlendMask: Top-Down Meets Bottom-Up for Instance Segmentation


Paper Code Results Date Stars