Mask RCNN

Extends Faster-RCNN

Problem: Faster-RCNN is not designed for pixel-to-pixel alignment between input and output.

Solution: Quantization-free layer that preserves spatial location (RoIAlign)

Architecture