Fig. 13From: Visual saliency detection for RGB-D images under a Bayesian frameworkEvaluation of the depth CNN learned using different training strategies on the NLPR dataset. a F-measure scores on the NLPR dataset for evaluation of the supervision transfer (ST) strategy and the fine-tuning (FT) strategy. b–f Qualitative comparison between the ST strategy and the FT strategy for 10k iterationsBack to article page