A Novel Saliency Detection Model based on Self Organizing Tree Algorithm

Abstract

Image saliency detection is very useful in many computer vision tasks while it still remains a challenging problem. In this paper, we propose a novel bottom-up model for visual saliency detection. The proposed model is based on both unsupervised learning approach and frequency domain analysis. Firstly, we incorporate the influence of center bias into our model, which is a common phenomenon that directs visual attention to the center of images in natural scenes. Hence, we introduce an unsupervised neural network that aims to measure the saliency center bias by exploring both color and texture low level cues. Secondly, the proposed model integrates Log-Gabor wavelets for visual saliency detection. This choice is justified by the fact that i) human visual system behavior detection of salient objects in a visual scene can be well modeled by band-pass filtering, ii) compared with the traditional model of receptive field, the Log-Gabor wavelets can better simulate the biological characteristics of the simple cortical cell in the receptive filed. Quantitative and qualitative experimental evaluation on MSRA-1000 public image dataset depicts the promising results from the proposed model by outperforming the relevant state of the art saliency detection models.