Pith. sign in

REVIEW 4 major objections 6 minor 43 references

Error Bounded Foreground and Background Modeling for Moving Object Detection in Satellite Videos

T0 review · 4 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Adding a bounded residual matrix to low-rank background/foreground decomposition lifts moving-object detection precision on satellite video.

desk verdict A modest, honest extension of LSD with an explicit residual term; the main flaw is that the ADMM schedule leaves the proven convergence region, so the reported numbers may describe the solver more than the model. read the letter →

arxiv 1908.09539 v1 pith:KDSFD5MA submitted 2019-08-26 cs.CV

Junpeng Zhang , Xiuping Jia , Jiankun Hu This is my paper · ORCID
classification cs.CV
keywords satellitevideomovingobjectdetectionbackgroundsubtractionlow-rankmatrixdecompositionstructuredsparsityresidualmodelingADMMsmalltarget
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Moving objects in satellite video cover only a few pixels and have low contrast, so simple background-subtraction models leave residuals that corrupt either the background or the foreground estimate. This paper proposes Extended Low-rank and Structured Sparse Decomposition (E-LSD), which adds a third matrix $E$ to the usual decomposition $D = B + S$, so that the low-rank background $B$ and structured-sparse foreground $S$ do not have to absorb data that fit neither. A bounded Frobenius-norm penalty on $E$ lets the model say 'unexplained' instead of forcing the residual into one of the two components. On two satellite videos, the paper reports higher detection precision and lower background rank than the original decomposition, together with faster convergence. The practical payoff is a cleaner background model for small, low-contrast moving targets, where earlier methods trade off detection accuracy against background purity.

What carries the argument

The central object is the three-matrix decomposition $D = B + S + E$, solved by the Alternating Direction Method of Multipliers. The nuclear norm $\|B\|_*$ relaxes the rank constraint, the structured sparsity-inducing norm $\|S\|_{\ell_1/\ell_\infty}$ penalizes foreground patterns whose nonzero pixels are scattered rather than clustered, and the squared Frobenius term $\|E\|_F^2$ bounds the residuals. The solver alternates singular value thresholding for $B$, a dual reformulation of the structured-sparsity update as a quadratic min-cost network-flow problem for $S$, and a closed-form update for $E$; the convergence claim rests on three-block ADMM analysis for a strongly convex block.

What would settle it

Run E-LSD on a synthetic video whose true background is a known low-rank matrix and whose foreground is a known set of small moving blobs, and compare the recovered $B$, $S$, $E$ against ground truth as the ADMM penalty is raised from inside to far outside the proven range; if the iterates stop at visibly wrong decompositions while smaller penalties succeed, the empirical penalty enlargement at the heart of the method is unsafe.

Watch

Extended reading notes

Core claim

The paper's central claim is that the failure of low-rank-plus-sparse decomposition on satellite video is not a weakness of the low-rank or the sparsity assumptions themselves but a missing model component: data that obey neither assumption. E-LSD therefore decomposes the observed frames as $D = B + S + E$ and minimizes the nuclear norm of $B$, a structured sparsity-inducing norm on $S$, and the squared Frobenius norm of $E$, subject to exact equality. The residual matrix $E$ absorbs transient artifacts such as noise, parallax, and illumination shifts, allowing $B$ to stay genuinely low-rank and $S$ to stay spatially coherent. Empirically the paper demonstrates on two satellite videos that E-LSD raises detection precision, lowers the rank of the estimated background from hundreds to tens, and reaches the same stopping criterion in fewer ADMM iterations than the original LSD. In the limit of an infinitely heavy penalty on $E$ the formulation reduces to LSD, so E-LSD is a strict generalization rather than a different model.

Load-bearing premise

The solver's proven convergence covers only a narrow range of penalty values, and the paper relies on the empirical premise that raising the penalty far beyond that range still lands on the true solution rather than a misleading decomposition.

Editorial extensions

If this is right

  • Low-contrast, few-pixel moving objects in satellite footage become detectable with fewer false alarms, because residuals no longer leak into the foreground mask.
  • The estimated background rank drops sharply (from hundreds to tens on both test videos), indicating a cleaner subspace for subsequent change detection and tracking.
  • E-LSD reaches the same relative reconstruction error in about 20 fewer iterations than LSD, cutting wall-clock time by roughly a factor of three in the reported experiments.
  • Since the model reduces to LSD when the residual penalty is large, any scene where LSD already works inherits the same behavior, making E-LSD a safe drop-in extension.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: a temporal smoothness prior on $E$ could separate slow illumination drift from moving targets, since $E$ is currently estimated frame-by-frame without temporal constraints.
  • Editorial inference: the same residual-absorption mechanism should help other small-target wide-area motion imagery, including high-altitude aerial video, whenever target contrast is near the noise floor.
  • Editorial inference: the reported sensitivity of background rank to the residual weight suggests a no-ground-truth tuning rule—scan the weight from large to small and stop where rank stops decreasing—which the paper does not state.
  • Editorial inference: the convergence advantage over LSD is likely to grow with residual magnitude, a prediction testable on synthetic data by injecting controlled noise levels and comparing iteration counts.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper addresses moving object detection in low-resolution satellite videos by extending the Low-rank and Structured Sparse Decomposition (LSD) model with an explicit residual term. The proposed E-LSD formulation decomposes the video matrix D into a low-rank background B, a structured-sparse foreground S, and a residual matrix E, with a squared-Frobenius penalty on E, and solves the resulting convex problem by a direct extension of ADMM. The authors claim that handling residuals explicitly yields cleaner background estimates, higher detection precision, and faster convergence than LSD, and they report experiments on two satellite videos (Video 001 and Video 002) comparing E-LSD with RPCA-PCP, GoDec, DECOLOR, and LSD.

Significance. If the central claims hold, the paper makes a useful practical contribution: it identifies a real limitation of low-rank-plus-structured-sparse decomposition on low-contrast satellite data and proposes a simple, convex remedy that is solvable with standard tools. The formulation is clearly motivated, the subproblems are standard (singular value thresholding and network-flow-based structured-sparsity projection), and the paper explicitly discusses parameter sensitivity. However, the validation is thin: only two videos from one satellite, no error bars or statistical tests, and the convergence of the proposed ADMM schedule is not supported by the cited theory. The significance is therefore conditional on closing the convergence gap and broadening the empirical evidence.

major comments (4)
  1. [Section II-C, Algorithm 1 and Eq. (20)] The convergence guarantee cited from [36] holds only for μ in (0, 7λ2/8), but Algorithm 1 initializes μ > 1 and increases it to μ × 1e5; with the experimental setting λ2 = λ1/5 and λ1 = 1/√p, the algorithm operates far outside the proven range. The stop criterion Eq. (20) checks only the primal residual ‖D−B−S−E‖_F/‖D‖_F, not stationarity. Consequently, the reported B, S, and E may not be minimizers of Eq. (4), and the rank and precision numbers in Table II could be properties of the penalty schedule rather than of the E-LSD model. Please provide a convergence proof for the actual μ schedule, or report dual residual/KKT-violation curves for the specific runs and verify that the reported solutions are stationary points.
  2. [Section II-B, Eqs. (3) and (4)] The bounded-error constraint in Eq. (3) is replaced by the penalized formulation in Eq. (4), but the paper does not establish an equivalence or a quantitative relationship between the parameter λ2 and the bound ζ; ζ never appears again in the algorithm or experiments. As written, E-LSD solves a penalized problem, not the constrained problem announced in the title and motivation. Please either state the exact relationship for which the two formulations coincide, or present E-LSD directly as a penalized model and adjust the claims accordingly.
  3. [Section III-C, Table II] The experimental validation is limited to two videos from the same satellite, with a single run per method and no error bars or statistical tests. The precision improvement on Video 001 is notable, but on Video 002 E-LSD has lower recall than LSD (79.6% vs. 82.2%) and essentially unchanged F1 (86.13% vs. 86.31%). The paper should temper the 'improved detection' claim, discuss the precision/recall trade-off explicitly, and ideally report variability across runs or additional satellite video sequences.
  4. [Section III-A and III-C, rank(B) as a background-quality metric] Low rank of the estimated background is used as a proxy for background modeling quality, but Table II shows that RPCA-PCP and GoDec achieve rank 2 with much worse detection performance, while E-LSD reports rank 65 and 76. Rank alone is therefore not a reliable indicator of background quality. The background-modeling improvement should be supported by a more direct measure, such as reconstruction error on known static regions, quantitative residual statistics, or a systematic visual evaluation, rather than rank and a single example.
minor comments (6)
  1. [Section III-B, text near Fig. 5] The text says 'As presented in Fig. 4, with fixed λ2, the detection performance by E-LSD increases', but Fig. 4 varies λ2 with fixed λ1; the λ1 effect is shown in Fig. 5. Please correct the cross-reference.
  2. [Section II-C.2, Eqs. (15)–(17)] The derivation of the dual problem assumes that 'all variables in h and ξ are assumed non-negative', but h = D − B − E + Y/μ is not non-negative in general. Please clarify the sign-handling transformation or provide the exact reduction to the network-flow formulation of [41] for arbitrary real-valued data.
  3. [Section II-C, Eq. (20)] Equation (20) uses B*, S*, and E* in the stopping condition although these are the quantities being estimated; the notation is confusing and should refer to the current iterates.
  4. [Section III-B] There is a typo: 'improved recall and prevision' should read 'improved recall and precision'.
  5. [Fig. 8] The column label 'GeDec' appears in the figure; the method name is spelled 'GoDec' elsewhere.
  6. [Section III-C] The comparison set is limited to methods from 2011–2015; since the paper claims state-of-the-art performance, it should at least discuss or compare with more recent satellite-video moving object detection approaches, including those cited in the introduction.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the E-LSD objective is independently defined and evaluated against external baselines.

full rationale

The paper's central claim is that solving the proposed objective (Eq. 4) yields improved detection precision and lower background rank compared with LSD and other baselines. The objective is defined from first principles as a three-term decomposition (nuclear norm, structured sparsity norm, and squared Frobenius residual) with a hard constraint D = B + S + E. This is not built from the evaluation metric (IoU-based precision/recall/F1) nor from the ground-truth bounding boxes. The performance improvement is measured against independent methods (RPCA-PCP, GoDec, DECOLOR, LSD) using a held-out test portion of the videos, with parameters selected on a separate 200-frame validation split. No fitted parameter is renamed as a prediction, and no cited result is invoked to forbid alternatives. The only notable weakness is an acknowledged gap between the theoretical ADMM convergence condition (mu in (0, 7*lambda2/8), per reference [36]) and the algorithm's empirical penalty schedule (mu grows to mu*1e5). The paper explicitly concedes that the theoretical analysis is conservative and that empirical enlargement leads to good solutions. This is a correctness/convergence risk, not circularity: the model and the solver are not defined in terms of the evaluation outcomes, so the reported comparisons are meaningful even if the solver's convergence is heuristic. The components inherited from prior work (LSD objective, ADMM, structured sparsity solvers) are standard scientific inheritance with independent support. Hence no circular step is present and the score is 0.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim rests on two domain priors (low-rank background, structured-sparse foreground), a modeling choice for residuals (Frobenius-bound), and an empirical convergence premise for the ADMM parameter schedule. No fundamentally new physical or mathematical entities are introduced.

free parameters (4)
  • lambda1 = 1/sqrt(p)
    Regularization weight for the structured sparsity term. Chosen by cross-validation on the first 200 frames of Video 001; the paper states 'lambda1 = 1/sqrt(p) ... were set for E-LSD'. The central decomposition depends on this value.
  • lambda2 = lambda1/5
    Regularization weight for the residual term. Also set by the cross-validation rule in Section III-B. The balance between foreground sparsity and residual size controls the rank of B.
  • batch length = length of the video (500 frames for evaluation)
    Section III-B sets the batch to the full video; the paper shows performance improves with batch length, so this choice affects the reported precision and recall.
  • binarization threshold = not specified
    The post-processing step ('utilize threshold segmentation as post-processing') needs a threshold to convert continuous foreground values to a mask; without it, the precision and recall numbers cannot be exactly reproduced.
assumptions (4)
  • domain assumption The background lies in a low-dimensional subspace (low rank/nuclear norm).
    Invoked in Eqs. (2)-(4); the paper assumes temporally stable background pixels form a low-rank matrix, and uses rank(B) as an evaluation proxy in Section III.
  • domain assumption Moving objects form spatially contiguous blobs that can be encoded by the l1/l-infinity norm over 3x3 pixel groups.
    Section II-A constructs G via 3x3 grid scanning; if true targets are subpixel or scattered, the structured sparsity prior could suppress real detections.
  • domain assumption The residuals that do not fit the background or foreground are well captured by a matrix E with small squared Frobenius norm.
    Eq. (4) replaces the inequality constraint with a Frobenius penalty; the paper does not justify that satellite-video residuals (sensor noise, atmospheric effects, parallax) are Frobenius-bounded.
  • standard math The direct extension of ADMM converges for the three-block problem under the practical mu schedule.
    Section II-C cites [36] for convergence when mu is in (0, 7*lambda2/8), then increases mu to mu_bar = mu*1e5. The convergence of the actual schedule is assumed empirically, not proven.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Error Bounded Foreground and Background Modeling for Moving Object Detection in Satellite Videos." pith.science (2026). https://pith.science/paper/KDSFD5MA

@misc{pith2026190809539,
  author       = {Pith},
  title        = {Pith review of: Error Bounded Foreground and Background Modeling for Moving Object Detection in Satellite Videos},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KDSFD5MA}},
  note         = {Machine review of arXiv:1908.09539}
}
read the original abstract

Detecting moving objects from ground-based videos is commonly achieved by using background subtraction techniques. Low-rank matrix decomposition inspires a set of state-of-the-art approaches for this task. It is integrated with structured sparsity regularization to achieve background subtraction in the developed method of Low-rank and Structured Sparse Decomposition (LSD). However, when this method is applied to satellite videos where spatial resolution is poor and targets' contrast to the background is low, its performance is limited as the data no longer fits adequately either the foreground structure or the background model. In this paper, we handle these unexplained data explicitly and address the moving target detection from space as one of the pioneer studies. We propose a technique by extending the decomposition formulation with bounded errors, named Extended Low-rank and Structured Sparse Decomposition (E-LSD). This formulation integrates low-rank background, structured sparse foreground and their residuals in a matrix decomposition problem. We provide an effective solution by introducing an alternative treatment and adopting the direct extension of Alternating Direction Method of Multipliers (ADMM). The proposed E-LSD was validated on two satellite videos, and experimental results demonstrate the improvement in background modeling with boosted moving object detection precision over state-of-the-art methods.

Figures

Figures reproduced from arXiv: 1908.09539 by the authors.

Figure 1
Figure 1. An exemplar on structured sparsity. (a) and (b) are two binary images [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Exemplar frames from Video 001 and Video 002. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Demonstration on the importance of introduced [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Performance evaluation with varying λ2 with λ1 = 0.0025 on Video 001. The ignorance of the residuals E in LSD would result in dropped performance in either moving object detection or background modeling. In LSD, the parameter λ balances the contribution of the relaxed …
Figure 5
Figure 5. Figure 5: Performance evaluation with different λ1 and λ2 on Video 001. The two curves in red and blue colors are exemplars with fixed λ1 = 0.001 and fixed λ2 = 0.005, respectively. −10 0 10 10−4 10−3 10−2 10−1 0 5 ·106 Bins λ2 Number of Pixels mean(E) [PITH_FULL_IMAGE:figures/…
Figure 6
Figure 6. Figure 6: Histogram of E with varying λ2 with fixed λ1 = 0.0025 on Video 001 which allows residuals to be encoded by the introduced E. As illustrated in Fig. 4c, with decreasing λ2, the rank of the estimated background decreases dramatically. At the same time, the precision of d…
Figure 7
Figure 7. Figure 7: Performance evaluation on batches of different lengths. Exemplar estimated backgrounds are provided at increasing batch lengths, where two sub-regions [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Compared with the original LSD approach, the E-LSD [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 8
Figure 8. Figure 8: Comparison with other existing state-of-the-art methods. Each column presents different approaches. The first row are the visualized detection results, [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Detection performance over varying IoU thresholds. [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]
Figure 10
Figure 10. Figure 10: Comparison on convergence between E-LSD and LSD approach. [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 39 canonical work pages

  1. [36]

    On the convergence of the direct extension of admm for three-block separable convex minimization models with one strongly convex function,

    X. Cai, D. Han, and X. Yuan, “On the convergence of the direct extension of admm for three-block separable convex minimization models with one strongly convex function,” Computational Optimization and Applications , vol. 66, no. 1, pp. 39–73, 2017

    work page 2017

  2. [1]

    Video satellite imagery super resolution via convolutional neural networks,

    Y . Luo, L. Zhou, S. Wang, and Z. Wang, “Video satellite imagery super resolution via convolutional neural networks,” IEEE Geoscience and Remote Sensing Letters , vol. 14, no. 12, pp. 2398–2402, 2017

    work page 2017

  3. [2]

    Planet application program interface: In space for life on earth. san francisco, ca,

    P. Team, “Planet application program interface: In space for life on earth. san francisco, ca,” 2016

    work page 2016

  4. [3]

    Small moving vehicle detection in a satellite video of an urban area,

    T. Yang, X. Wang, B. Yao, J. Li, Y . Zhang, Z. He, and W. Duan, “Small moving vehicle detection in a satellite video of an urban area,” Sensors, vol. 16, no. 9, p. 1528, 2016

    work page 2016

  5. [4]

    Low-rank decomposition and total variation regularization of hyperspectral video sequences,

    Y . Xu, Z. Wu, J. Chanussot, M. Dalla Mura, A. L. Bertozzi, and Z. Wei, “Low-rank decomposition and total variation regularization of hyperspectral video sequences,” IEEE Transactions on Geoscience and Remote Sensing , vol. 56, no. 3, pp. 1680–1694, 2017

    work page 2017

  6. [5]

    Spatiotemporal scene interpretation of space videos via deep neural network and tracklet analysis,

    L. Mou and X. X. Zhu, “Spatiotemporal scene interpretation of space videos via deep neural network and tracklet analysis,” in Geoscience and Remote Sensing Symposium (IGARSS), 2016 IEEE International . IEEE, 2016, pp. 1823–1826

    work page 2016

  7. [6]

    Object tracking in satellite videos by fusing the kernel correlation filter and the three-frame- difference algorithm,

    B. Du, Y . Sun, S. Cai, C. Wu, and Q. Du, “Object tracking in satellite videos by fusing the kernel correlation filter and the three-frame- difference algorithm,” IEEE Geoscience and Remote Sensing Letters , vol. 15, no. 2, pp. 168–172, 2018

    work page 2018

  8. [7]

    Satellite multi-vehicle tracking under inconsistent detection conditions by bilevel k-shortest paths opti- mization,

    J. Zhang, X. Jia, J. Hu, and K. Tan, “Satellite multi-vehicle tracking under inconsistent detection conditions by bilevel k-shortest paths opti- mization,” in 2018 Digital Image Computing: Techniques and Applica- tions (DICTA). IEEE, 2018, pp. 1–8

    work page 2018

Show all 43 references
  1. [8]

    Tracking in aerial hyperspectral videos using deep kernelized correlation filters,

    B. Uzkent, A. Rangnekar, and M. J. Hoffman, “Tracking in aerial hyperspectral videos using deep kernelized correlation filters,” IEEE Transactions on Geoscience and Remote Sensing, no. 99, pp. 1–13, 2018

    2018

  2. [9]

    Vehicle detection and traffic density monitoring from very high resolution satellite video data,

    G. Kopsiaftis and K. Karantzalos, “Vehicle detection and traffic density monitoring from very high resolution satellite video data,” in Geoscience and Remote Sensing Symposium (IGARSS), 2015 IEEE International . IEEE, 2015, pp. 1881–1884

    2015

  3. [10]

    Background subtraction techniques: a review,

    M. Piccardi, “Background subtraction techniques: a review,” in 2004 IEEE International Conference on Systems, Man and Cybernetics (IEEE Cat. No. 04CH37583) , vol. 4. IEEE, 2004, pp. 3099–3104

    2004

  4. [11]

    Background subtraction for automated multisensor surveillance: a comprehensive review,

    M. Cristani, M. Farenzena, D. Bloisi, and V . Murino, “Background subtraction for automated multisensor surveillance: a comprehensive review,”EURASIP Journal on Advances in Signal Processing , vol. 2010, no. 1, p. 343057, 2010

    2010

  5. [12]

    A comprehensive review of background subtraction algorithms evaluated with synthetic and real videos,

    A. Sobral and A. Vacavant, “A comprehensive review of background subtraction algorithms evaluated with synthetic and real videos,” Com- puter Vision and Image Understanding , vol. 122, pp. 4–21, 2014

    2014

  6. [13]

    Robust pca via principal component pursuit: A review for a comparative evaluation in video surveillance,

    T. Bouwmans and E. H. Zahzah, “Robust pca via principal component pursuit: A review for a comparative evaluation in video surveillance,” Computer Vision and Image Understanding , vol. 122, pp. 22–34, 2014

    2014

  7. [14]

    Decomposition into low-rank plus additive matrices for back- ground/foreground separation: A review for a comparative evaluation with a large-scale dataset,

    T. Bouwmans, A. Sobral, S. Javed, S. K. Jung, and E.-H. Za- hzah, “Decomposition into low-rank plus additive matrices for back- ground/foreground separation: A review for a comparative evaluation with a large-scale dataset,” Computer Science Review , vol. 23, pp. 1– 71, 2017

    2017

  8. [15]

    On the applications of robust pca in image and video processing,

    T. Bouwmans, S. Javed, H. Zhang, Z. Lin, and R. Otazo, “On the applications of robust pca in image and video processing,” Proceedings of the IEEE , vol. 106, no. 8, pp. 1427–1457, 2018

    2018

  9. [16]

    Principal component analysis,

    I. Jolliffe, “Principal component analysis,” in International Encyclopedia of Statistical Science . Springer, 2011, pp. 1094–1096

    2011

  10. [17]

    Robust principal component analysis: Exact recovery of corrupted low-rank matrices via convex optimization,

    J. Wright, A. Ganesh, S. Rao, Y . Peng, and Y . Ma, “Robust principal component analysis: Exact recovery of corrupted low-rank matrices via convex optimization,” in Advances in Neural Information Processing Systems, 2009, pp. 2080–2088

    2009

  11. [18]

    Linearized alternating direction method with adaptive penalty for low-rank representation,

    Z. Lin, R. Liu, and Z. Su, “Linearized alternating direction method with adaptive penalty for low-rank representation,” in Advances in Neural Information Processing Systems , 2011, pp. 612–620

    2011

  12. [19]

    Robust principal component analysis?

    E. J. Cand `es, X. Li, Y . Ma, and J. Wright, “Robust principal component analysis?” Journal of the ACM (JACM) , vol. 58, no. 3, p. 11, 2011

    2011

  13. [20]

    Godec: Randomized low-rank & sparse matrix decomposition in noisy case,

    T. Zhou and D. Tao, “Godec: Randomized low-rank & sparse matrix decomposition in noisy case,” in International Conference on Machine Learning. Omnipress, 2011

    2011

  14. [21]

    Rank-sparsity incoherence for matrix decomposition,

    V . Chandrasekaran, S. Sanghavi, P. A. Parrilo, and A. S. Willsky, “Rank-sparsity incoherence for matrix decomposition,” SIAM Journal on Optimization , vol. 21, no. 2, pp. 572–596, 2011

    2011

  15. [22]

    Foreground detection based on low-rank and block-sparse matrix decomposition,

    C. Guyon, T. Bouwmans, and E.-H. Zahzah, “Foreground detection based on low-rank and block-sparse matrix decomposition,” in2012 19th IEEE International Conference on Image Processing . IEEE, 2012, pp. 1225–1228. ERROR BOUNDED FOREGROUND AND BACKGROUND MODELING FOR MOVING OBJEC...

    2012

  16. [23]

    Or- pca with mrf for robust foreground detection in highly dynamic back- grounds,

    S. Javed, S. H. Oh, A. Sobral, T. Bouwmans, and S. K. Jung, “Or- pca with mrf for robust foreground detection in highly dynamic back- grounds,” in Asian Conference on Computer Vision . Springer, 2014, pp. 284–299

    2014

  17. [24]

    Moving object detection by detecting contiguous outliers in the low-rank representation,

    X. Zhou, C. Yang, and W. Yu, “Moving object detection by detecting contiguous outliers in the low-rank representation,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 35, no. 3, pp. 597– 610, 2013

    2013

  18. [25]

    Corola: a sequential solution to moving object detection using low-rank approximation,

    M. Shakeri and H. Zhang, “Corola: a sequential solution to moving object detection using low-rank approximation,” Computer Vision and Image Understanding, vol. 146, pp. 27–39, 2016

    2016

  19. [26]

    Structured variable selection with sparsity-inducing norms,

    R. Jenatton, J.-Y . Audibert, and F. Bach, “Structured variable selection with sparsity-inducing norms,” Journal of Machine Learning Research , vol. 12, no. Oct, pp. 2777–2824, 2011

    2011

  20. [27]

    Proximal methods for sparse hierarchical dictionary learning

    R. Jenatton, J. Mairal, G. Obozinski, and F. R. Bach, “Proximal methods for sparse hierarchical dictionary learning.” in ICML, vol. 1. Citeseer, 2010, p. 2

    2010

  21. [28]

    Robust and practical face recognition via structured sparsity,

    K. Jia, T.-H. Chan, and Y . Ma, “Robust and practical face recognition via structured sparsity,” in European Conference on Computer Vision . Springer, 2012, pp. 331–344

    2012

  22. [29]

    Background subtraction based on low-rank and structured sparse decomposition,

    X. Liu, G. Zhao, J. Yao, and C. Qi, “Background subtraction based on low-rank and structured sparse decomposition,” IEEE Transactions on Image Processing, vol. 24, no. 8, pp. 2502–2514, 2015

    2015

  23. [30]

    Model selection and estimation in regression with grouped variables,

    M. Yuan and Y . Lin, “Model selection and estimation in regression with grouped variables,” Journal of the Royal Statistical Society: Series B (Statistical Methodology) , vol. 68, no. 1, pp. 49–67, 2006

    2006

  24. [31]

    Proximal methods for hierarchical sparse coding,

    R. Jenatton, J. Mairal, G. Obozinski, and F. Bach, “Proximal methods for hierarchical sparse coding,” Journal of Machine Learning Research , vol. 12, no. Jul, pp. 2297–2334, 2011

    2011

  25. [32]

    Background subtraction via generalized fused lasso foreground modeling,

    B. Xin, Y . Tian, Y . Wang, and W. Gao, “Background subtraction via generalized fused lasso foreground modeling,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2015, pp. 4676–4684

    2015

  26. [33]

    Stable principal com- ponent pursuit,

    Z. Zhou, X. Li, J. Wright, E. Candes, and Y . Ma, “Stable principal com- ponent pursuit,” in 2010 IEEE International Symposium on Information Theory. IEEE, 2010, pp. 1518–1522

    2010

  27. [34]

    Simultaneous video stabilization and moving object detection in turbulence,

    O. Oreifej, X. Li, and M. Shah, “Simultaneous video stabilization and moving object detection in turbulence,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 35, no. 2, pp. 450–462, 2013

    2013

  28. [35]

    The direct extension of admm for multi-block convex minimization problems is not necessarily conver- gent,

    C. Chen, B. He, Y . Ye, and X. Yuan, “The direct extension of admm for multi-block convex minimization problems is not necessarily conver- gent,” Mathematical Programming, vol. 155, no. 1-2, pp. 57–79, 2016

    2016

  29. [37]

    Distributed optimization and statistical learning via the alternating direction method of multipliers,

    S. Boyd, N. Parikh, E. Chu, B. Peleato, J. Eckstein et al. , “Distributed optimization and statistical learning via the alternating direction method of multipliers,” F oundations and Trends R© in Machine Learning , vol. 3, no. 1, pp. 1–122, 2011

    2011

  30. [38]

    A singular value thresholding al- gorithm for matrix completion,

    J.-F. Cai, E. J. Cand `es, and Z. Shen, “A singular value thresholding al- gorithm for matrix completion,” SIAM Journal on Optimization , vol. 20, no. 4, pp. 1956–1982, 2010

    1956

  31. [39]

    Boyd and L

    S. Boyd and L. Vandenberghe, Convex optimization . Cambridge university press, 2004

    2004

  32. [40]

    Lagrange multipliers revisited,

    M. Slater, “Lagrange multipliers revisited,” Cowles Foundation for Research in Economics, Yale University, Tech. Rep., 1959

    1959

  33. [41]

    Network flow algorithms for structured sparsity,

    J. Mairal, R. Jenatton, F. R. Bach, and G. R. Obozinski, “Network flow algorithms for structured sparsity,” in Advances in Neural Information Processing Systems, 2010, pp. 1558–1566

    2010

  34. [42]

    The augmented lagrange multiplier method for exact recovery of corrupted low-rank matrices,

    Z. Lin, M. Chen, and Y . Ma, “The augmented lagrange multiplier method for exact recovery of corrupted low-rank matrices,” arXiv preprint arXiv:1009.5055, 2010

    2010 arXiv

  35. [43]

    Block-sparse rpca for consistent foreground detection,

    Z. Gao, L.-F. Cheong, and M. Shan, “Block-sparse rpca for consistent foreground detection,” in European Conference on Computer Vision . Springer, 2012, pp. 690–703. Junpeng Zhang received the B.Sci. degree from the China University of Mining and Technology, Xuzhou, China, in 2...

    1999

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.