Name
Abstract | ||
|---|---|---|
Bottom-up saliency detection aims to detect salient areas within natural images usually without learning from labeled images. Typically, the saliency map of an image is inferred by only using the information within this image (referred to as the “current image”). While efficient, such single-image-based methods may fail to obtain reliable results, because the information within a single image may be insufficient for defining saliency. In this paper, we investigate how saliency detection can benefit from the nearest neighbor structure in the image space. First, we show that existing methods can be improved by extending them to include the visual neighborhood information. This verifies the significance of the neighbors. Next, a solution of multitask sparsity pursuit is proposed to integrate the current image and its neighbors to collaboratively detect saliency. The integration is done by first representing each image as a feature matrix, and then seeking the consistently sparse elements from the joint decompositions of multiple matrices into pairs of low-rank and sparse matrices. The computational procedure is formulated as a constrained nuclear norm and $\ell_{2,1}$-norm minimization problem, which is convex and can be solved efficiently with the augmented Lagrange multiplier method. Besides the nearest neighbor structure in the visual feature space, the proposed model can also be generalized to handle multiple visual features. Extensive experiments have clearly validated its superiority over other state-of-the-art methods. |
Year | DOI | Venue |
|---|---|---|
2013 | 10.1109/TCSVT.2013.2248495 | IEEE Transactions on Circuits and Systems for Video Technology |
Keywords | Field | DocType |
multimodal modeling,image representation,sparse elements,multitask learning,natural images,labeled images,sparse matrices,single-image-based methods,saliency map,convex programming,bottom-up saliency detection,saliency detection,augmented lagrange multiplier method,feature matrix,sparsity and low-rankness,visual attention,nearest neighbor structure,1-norm minimization problem,visual neighborhood information,ℓ2,visual feature space,joint decompositions,image space,constrained nuclear norm,reliability,visualization,matrix decomposition,computational modeling | k-nearest neighbors algorithm,Computer vision,Feature vector,Feature detection (computer vision),Pattern recognition,Computer science,Matrix (mathematics),Salience (neuroscience),Matrix norm,Artificial intelligence,Convex optimization,Sparse matrix | Journal |
Volume | Issue | ISSN |
23 | 6 | 1051-8215 |
Citations | PageRank | References |
6 | 0.47 | 29 |
Authors | ||
6 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Congyan Lang | 1 | 353 | 39.20 |
| Jiashi Feng | 2 | 2165 | 140.81 |
| Guangcan Liu | 3 | 2515 | 76.85 |
| Jinhui Tang | 4 | 5180 | 212.18 |
| Shuicheng Yan | 5 | 767 | 25.71 |
| Jiebo Luo | 6 | 6314 | 374.00 |