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

Table 2

The prediction performances of models A, B, and C, as well as our proposed autoencoder.

Datasets	Criteria	Model A	Model B	Model C	Proposed
NUS	CC	0.5220	0.5227	0.5097	0.5310
	KLDiv	1.2538	1.3408	1.5606	1.2323
	AUC	0.8353	0.8351	0.7841	0.8501
	NSS	2.1198	2.1727	2.1301	2.1195
NCTU	CC	0.7607	0.8043	0.7967	0.8034
	KLDiv	0.3900	0.4152	0.3869	0.3593
	AUC	0.8552	0.8641	0.8618	0.8671
	NSS	1.7348	1.8914	1.8227	1.8405