AN ENHANCED FUSION BASED REGION SELECTION FOR CO-SALIENT DETECTION

Computer Engineering

Human eye is perceptually more sensitive to certain colors and intensities and objects with such features are considered more salient. Co-saliency is a term for common visual attention point. Co-saliency detection aims at discovering the common salient objects existing in multiple images, which is a relatively under-explored area. Most existing methods combine multiple saliency cues based on fixed weights, and ignore the intrinsic relationship of these cues. In this paper, it proposes a method for finding salient region for single image, and co-salient region for pair of images as well as multiple images. And it provide a general saliency map fusion framework, which exploits the relationship of multiple saliency cues and obtains the self-adaptive weight to generate the final saliency/co-saliency map. Previous methods sometimes inaccurately predict the background as common salient object in case of complex images. The feature representation of the co-salient regions should be both similar and consistent. Therefore, the matrix jointing these feature histograms appears low rank. By considering the low-rank matrix of features, which defines the independence of the features of the co-salient region, it can find energy. This energy will further use to determine the co-saliency map. It helps to determine the co-salient map for complex images efficiently. Detection of Co-salient image regions is useful in applications such as co-segmentation, common pattern discovery, co-recognition, Image retrieval etc.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.