An improved saliency detection using wavelet transform

With a saliency map providing the information of locations where are visually salient to human visual system, region based image processing can be performed more efficiently. In this paper, we propose a method to compute the saliency map. Firstly we reconstruct from different levels in wavelet transform with different weight to obtain feature maps which can represent features from edge to texture to gain the local saliency map. Then we segment and merge similar regions to calculate the global saliency. The final saliency map is computed by modulating local saliency value with its global saliency value. Our proposed model outperforms many of the relevant state-of-the-art saliency detection models.