The learning of "Saliency Detection via Dense and Sparse Reconstruction" (L.-H. Zhang, X. Ruan, M.-H. Yang. IEEE International Conference on Computer Vision, Dec, 2013, pp. 2976-2983)
Use image boundaries via superpixels as background templates, from which dense and sparse appearance models are constructed. Improved saliency map by error propagation, object-biased Gaussian model, and use Bayesian inference integrate final saliency map.
- Model overview
- Generate superpixels using the simple linear iterative clustering (SLIC) algorithm, and extract the D-dimensional feature of each boundary segment and construct the background template set.
- Use dense reconstruction errors to measure the saliency of each region via Principal Component Analysis.
- Use sparse reconstruction error to measure the saliency of each region via sparse representation.
- Apply the K-means algorithm to cluster N image segments, then compare segment i between the other segments belonging to cluster k to smooth the reconstruction errors generated by dense and sparse appearance models.
- Utilize the similarity between pixel z and its corresponding segment n at scale s as the weight to average the multi-scale reconstruction errors, and get pixel-level saliency.
- Use an object-biased Gaussian model to refine saliency map.
- Take one saliency map as the prior and use the other one instead of Lab color information to compute the likelihoods, which integrates more diverse information from different saliency maps.
- Use these two posterior probabilities to compute an integrated saliency map.
- Specific description and supplementary knowledge
- Dense Reconstruction Error
Background templates: B=b1,b2…bM B∈RD×M segment i (i∈[1,N] ),
Principal Component Analysis (PCA):
A multivariate statistical analysis method in which multiple variables are selected by linear transformation to select fewer significant variables.
In this paper, use PCA to analysis the background templates. Get a more concise expression of the background, and to some extent filter out the interference that does not belong to the background.
- Construct covariance matrices of background samples
C=x1-x',x2-x'…xn-x'x1-x',x2-x'…xn-x'T
- Eigendecomposition
C=UBΛUB-1
Matlab: [pc, latent, explained] = pcacov (COVM);
pc=UB latent=Λ , explained is percentage
Select 95% of the ingredients as the main ingredients, and constitute newUB .
- Compute the reconstruction coefficient of segment i
βi=UBTxi-x'
- and the dense reconstruction error of segment i is
εid=xi-UBβi+x'22
-
- Object-biased Gaussian model
- Object center (The expectation of salient position coordinates)
x0=iE(i)jE(j)xi y0=iE(i)jE(j)yi
- Gaussian model
Gz=exp-xz-ux22σx2+yz-uy22σy2
ux=x0, uy=y0
Different saliency maps have different Object biases.
- The saliency of pixel z is
Sz=G0z*Ez