Deep Image Understanding Using Multilayered Contexts

Table 1

Model performance with and without context information.

Model	SGGen		Meteor
	R@50	R@100	mAP
Baseline	10.71	12.96	5.30
Baseline + RCN	12.21	14.57	5.61
Baseline + CCN	10.90	13.12	5.62
Baseline + RCN + CCN	12.66	14.85	5.71