Continuous Adaptation for Interactive Object Segmentation by Learning from Corrections

Open paper

Benchmark	Model	Rank	Results
interactive-segmentation-on-berkeley	IA+SA	–	NoC@90: 4.94
interactive-segmentation-on-davis	IA+SA	–	NoC@85: 5.16
interactive-segmentation-on-grabcut	IA+SA	–	NoC@90: 3.07