Salient object detection based on objectness

We propose a novel salient object detection method based on objectness by generating salient object bounding box proposals. Smooth images of edge-preserving well and smooth background are obtained by using L₀ Gradient Minimization. After oversegmenting smooth images using SLIC algorithm, candidate bounding boxes as well as center and neighbor areas in a bounding box are then determined. Combing objectness scores, contrast, center and background priors, ranking bounding boxes containing salient objects are obtained finally. Although saliency map can't directly be obtained by our algorithm unlike others salient object detection methods, salient objects containing meaningful context information can be labeled with ranking bounding boxes. Experiment results demonstrate that the proposed algorithm can yield better detection and run faster than salient object detection of SVO method. Based on internal connection between saliency and objectness, algorithm proposed in this paper can be viewed as a novel try on salient object detection.