Joint Dictionary Learning for Multispectral Change Detection
Citations
1,001 citations
Cites methods from "Joint Dictionary Learning for Multi..."
...Typical unsupervised feature learning methods include, but not limited to, principal component analysis, k-means clustering, sparse coding [26]–[28], [33], [44] and autoencoder [29], [31]....
[...]
550 citations
462 citations
254 citations
231 citations
References
9,950 citations
"Joint Dictionary Learning for Multi..." refers background in this paper
...Because (2) is NP-hard in general, some relaxation strategies, including basis pursuit [47] and orthogonal matching pursuit [48], are exploited to approximate the solution of (2)....
[...]
9,380 citations
"Joint Dictionary Learning for Multi..." refers background or methods in this paper
...Because (2) is NP-hard in general, some relaxation strategies, including basis pursuit [47] and orthogonal matching pursuit [48], are exploited to approximate the solution of (2)....
[...]
...Broadly, the techniques to the selection of dictionary can be categorized into two families: 1) transformation-based methods [48] and 2) learning-based...
[...]
4,387 citations
3,361 citations
2,731 citations
"Joint Dictionary Learning for Multi..." refers background or methods in this paper
...The constraint in (9) prevents the dictionary D from being large, which will lead to arbitrarily small values of sparse coefficients [42]....
[...]
...In each feature-sign step [42], [46], [53], the analytical solution ŝnew i is calculated by (12) under the current active set and signs....
[...]
...According to the feature-sign search method proposed in [42] and [53], the subdifferential of the (12) is discussed in the situation for different values of the coefficient s j) i ....
[...]
...Sparsity constraint has shown promising results in finding succinct representations of stimuli [42]....
[...]
...In this paper, a Lagrange dual method [42] is adopted to compute the dictionary D....
[...]