Saliency Detection via Sparse Reconstruction Errors of Covariance Descriptors on Riemannian Manifolds

Zhang, Yingying; Wang, Zhipeng; Lv, Xiaodong

doi:10.1007/s00034-016-0267-x

Cited by 10 publications

(10 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proposed approach shares similarities with previous methods mainly including [24]， [25]. The model in [24] only considers some basic vector features for sparse representation but not the covariance matrices as in the proposed model.…”

Section: Related Workmentioning

confidence: 95%

“…In addition, it is worth noticing that incorporating high level prior into saliency estimation has been previously investigated by a number of researchers. For example, some latest methods [19,[23][24][25][26] exploit boundary priors, which indicated that regions along the image boundary are more likely to be the background. The boundary prior is proved to be more general than the center prior [27].…”

Section: Introductionmentioning

confidence: 99%

“…Despite the differences in these methods, they all use the vector feature to compute the saliency measure except [25]. However, the vector feature considers features equally without considering the relationship of these different features, thereby it becomes invalid when there exit complex textures or various colors in the salient regions.…”

Section: Introductionmentioning

confidence: 99%

“…Region covariance matrices naturally provide nonlinear integration of different features by modeling their correlations and capture local image structures better than standard vector features [29]. Our previous work in [25] has exploited sparse coding of covariance matrices for salient region detection. However, using the reconstruction error of background dictionary directly to measure saliency may produce unsatisfying results in complex scenes and influence the overall performance of the method.…”

Section: Introductionmentioning

confidence: 99%

“…We consider the background-based mode to get the rough saliency estimation and (a) (b) (c) (d) (e) provide the foreground dictionary for the foreground-based saliency mode. After obtaining the foreground-based saliency result, we propose some refinements to overcome the disadvantage of the corresponding the foreground-based saliency result and develop the advantage, which is different from the refinements for the corresponding the background-based saliency result [25] and the refinements for the integrated result [28]. Fig.1 shows some saliency results generated by the proposed method.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Two-stage sparse coding of region covariance via Log-Euclidean kernels to detect saliency

2017

Self Cite

View full text Add to dashboard Cite

In this paper, we present a novel bottom-up saliency detection algorithm from the perspective of covariance matrices on a Riemannian manifold. Each superpixel is described by a region covariance matrix on Riemannian Manifolds. We carry out a two-stage sparse coding scheme via Log-Euclidean kernels to extract salient objects efficiently. In the first stage, given background dictionary on image borders, sparse coding of each region covariance via Log-Euclidean kernels is performed. The reconstruction error on the background dictionary is regarded as the initial saliency of each superpixel. In the second stage, an improvement of the initial result is achieved by calculating reconstruction errors of the superpixels on foreground dictionary, which is extracted from the first stage saliency map. The sparse coding in the second stage is similar to the first stage, but is able to effectively highlight the salient objects uniformly from the background. Finally, three post-processing methods-highlight-inhibition function, context-based saliency weighting, and the graph cut-are adopted to further refine the saliency map. Experiments on four public benchmark datasets show that the proposed algorithm outperforms the state-of-the-art methods in terms of precision, recall and mean absolute error, and demonstrate the robustness and efficiency of the proposed method.

show abstract

Section: Related Workmentioning

confidence: 95%