Deep Multimodal Fusion Autoencoder for Saliency Prediction of RGB-D Images

Table 1

The evaluation results of various saliency models.

Datasets	Criteria	Itti	GBVS	QFT	Fang	Qi	DeepFix	ML-net	DVA	Proposed
NUS	CC	0.341	0.396	0.163	0.333	0.371	0.4322	0.446	0.4549	0.5310
	KLDiv	1.457	1.374	1.795	1.560	1.505	1.8138	1.780	2.4349	1.2323
	AUC	0.788	0.824	0.682	0.795	0.806	0.7699	0.766	0.7236	0.8501
	NSS	1.236	1.441	0.568	1.209	1.357	1.6608	1.821	1.7962	2.1195
NCTU	CC	0.449	0.533	0.292	0.542	0.595	0.7974	0.696	0.6834	0.8034
	KLDiv	0.738	0.619	0.893	0.674	0.616	1.3083	0.900	1.1045	0.3593
	AUC	0.753	0.789	0.698	0.806	0.816	0.8650	0.835	0.8035	0.8671
	NSS	0.978	1.184	0.695	1.264	1.373	1.8575	1.588	1.5546	1.8405