In this paper, an image saliency detection method based on regional label fusion is proposed to solve the problems with fuzzy boundaries, unclear profile, and less interior density commonly existing in the researches of salient region detection. The image is segmented by super pixel segmentation algorithm, then the spectral clustering is carried out for the super pixel region to reduce the number of regions, thereby the label set with the boundary information could be obtained. Next, three salient features of the image have been fused under conditional random field model to generate the coarse saliency map. Afterwards, regional label fusion method is operated, which organically fuses the boundary information into the coarse saliency map by using the salient mean value calculated with the label information as the regional salient features, moreover, together with adaptive threshold segmentation algorithm to acquire reconstructed saliency map. At last, accurate salient region detection is achieved by calculating with a tag indicating vector defined and reconstructed coarse saliency map. Experimental results show that the salient regions obtained by this algorithm display clearer boundary contours and that the density of salient regions has been greatly improved compared with the other six significant detection methods prevailed in recent years.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.