Automatically extracting salient regions in natural images

Abstract

Salient regions extraction plays an important role in image analysis and processing, and it is widely used in image compression, coding, content-based image retrieve and so on. The use of salient regions can improve the efficiency of image processing and reduce the computational complexity. An approach for salient regions extraction in natural image is proposed in this paper. By analyzing the phase spectra of different color feature maps in different scales in frequency domain, the corresponding saliency map is constructed in spatial domain. Based on the saliency map and visual attention mechanism, the salient regions in an image are extracted. This approach has been tested on many natural images. Experiment results and quantitative and qualitative analysis have been presented in this paper. The proposed results show that this algorithm can extract salient regions consistent with human vision and has effective results in compution speed and antinoise ability.