Semi-Supervised Image Classification by Nonnegative Sparse Neighborhood Propagation

Zhang, Zhao; Zhang, Li; Zhao, Mingbo; Jiang, Weizhong; Liang, Yuchen; Li, Fanzhang

doi:10.1145/2671188.2749292

Cited by 35 publications

(28 citation statements)

References 26 publications

(28 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(2) To make the coding process robust to noise and outliers in data, and potentially reduce the factorization and reconstruction errors, J-RFDL uses the sparse and robust L 2, 1 -norm [35][36][37].…”

Section: Introductionmentioning

confidence: 99%

Learning Hybrid Representation by Robust Dictionary Learning in Factorized Compressed Space

Ren

Zhang

et al. 2020

IEEE Trans. on Image Process.

Self Cite

View full text Add to dashboard Cite

In this paper, we investigate the robust dictionary learning (DL) to discover the hybrid salient low-rank and sparse representation in a factorized compressed space. A Joint Robust Factorization and Projective Dictionary Learning (J-RFDL) model is presented. The setting of J-RFDL aims at improving the data representations by enhancing the robustness to outliers and noise in data, encoding the reconstruction error more accurately and obtaining hybrid salient coefficients with accurate reconstruction ability. Specifically, J-RFDL performs the robust representation by DL in a factorized compressed space to eliminate the negative effects of noise and outliers on the results, which can also make the DL process efficient. To make the encoding process robust to noise in data, J-RFDL clearly uses sparse L2, 1-norm that can potentially minimize the factorization and reconstruction errors jointly by forcing rows of the reconstruction errors to be zeros. To deliver salient coefficients with good structures to reconstruct given data well, J-RFDL imposes the joint low-rank and sparse constraints on the embedded coefficients with a synthesis dictionary. Based on the hybrid salient coefficients, we also extend J-RFDL for the joint classification and propose a discriminative J-RFDL model, which can improve the discriminating abilities of learnt coefficients by minimizing the classification error jointly. Extensive experiments on public datasets demonstrate that our formulations can deliver superior performance over other state-of-the-art methods. Index Terms-Hybrid salient representation, robust factorized compression, robust projective dictionary learning, classificationW ---------------- J. Ren is with the

show abstract

Section: Introductionmentioning

confidence: 99%

Learning Hybrid Representation by Robust Dictionary Learning in Factorized Compressed Space

Ren

Zhang

et al. 2020

IEEE Trans. on Image Process.

Self Cite

View full text Add to dashboard Cite

show abstract

“…As a classical graph-based SSC algorithm [61][62][63][64][65][66][67][68][69][70], label propagation (LP) [1][2][3][4][5][6][7][8][9] has been widely utilized due to its effectiveness and efficiency. LP models predict the labels of samples by propagating label information from labeled data to unlabeled data using their geometric structures and initial state [1][2][3][4][5][6][7][8][9], namely, by balancing the label fitness and the manifold smoothness.…”

Section: Introductionmentioning

confidence: 99%

“…Representative transductive LP algorithms mainly include Gaussian fields and harmonic function (GFHF) [1], linear neighborhood propagation (LNP) [2], special LP (SLP) [3], learning with local and global Consistency (LLGC) [5], projective label propagation (ProjLP) via label embedding [6], prior-class-dissimilarity-based LNP (CD-LNP) [4], sparse neighborhood propagation (SparseNP) [8], positive and negative label propagation (PN-LP) [7], and adaptive neighborhood propagation (AdaptiveNP) [9]. GFHF, LLGC, LNP and SLP optimize similar objective functions to predict the labels of samples by receiving information partly from the neighborhood and partly from the initial state.…”

Section: Introductionmentioning

confidence: 99%

“…Thus, the labels that are determined over the original predicted soft labels may be inaccurate or even completely incorrect; LLGC, LNP, GFHF, SLP, CD-LNP, PN-LP and AdaptiveNP suffer from this problem. To eliminate the negative effects of noise and mixed signs in the estimated soft labels, SparseNP [8] and ProjLP [6] were introduced as feasible approaches. ProjLP can predict more discriminating "deep" soft labels of data by using label embedding to reduce the noise and mixed signs in soft labels to enhance the classification performance [6].…”

Section: Introductionmentioning

confidence: 99%

“…ProjLP can predict more discriminating "deep" soft labels of data by using label embedding to reduce the noise and mixed signs in soft labels to enhance the classification performance [6]. In contrast to ProjLP, SparseNP explicitly employs a sparse regularization on the predicted soft labels for the joint optimization to reduce the unfavorable mixed signs and the noise in the soft labels on the classification result [8]. In this paper, we develop another method and propose a simple yet effective approach for removing noise and mixed signs from the predicted soft labels via clean label recovery.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Robust Triple-Matrix-Recovery-Based Auto-Weighted Label Propagation for Classification

Zhang

Zhao

et al. 2020

IEEE Trans. Neural Netw. Learning Syst.

Self Cite

View full text Add to dashboard Cite

The graph-based semi-supervised label propagation algorithm has delivered impressive classification results. However, the estimated soft labels typically contain mixed signs and noise, which cause inaccurate predictions due to the lack of suitable constraints. Moreover, available methods typically calculate the weights and estimate the labels in the original input space, which typically contains noise and corruption. Thus, the encoded similarities and manifold smoothness may be inaccurate for label estimation. In this paper, we present effective schemes for resolving these issues and propose a novel and robust semi-supervised classification algorithm, namely, the triple-matrix-recovery-based robust auto-weighted label propagation framework (ALP-TMR). Our ALP-TMR introduces a triple matrix recovery mechanism to remove noise or mixed signs from the estimated soft labels and improve the robustness to noise and outliers in the steps of assigning weights and predicting the labels simultaneously. Our method can jointly recover the underlying clean data, clean labels and clean weighting spaces by decomposing the original data, predicted soft labels or weights into a clean part plus an error part by fitting noise. In addition, ALP-TMR integrates the auto-weighting process by minimizing reconstruction errors over the recovered clean data and clean soft labels, which can encode the weights more accurately to improve both data representation and classification. By classifying samples in the recovered clean label and weight spaces, one can potentially improve the label prediction results. The results of extensive experiments demonstrated the satisfactory performance of our ALP-TMR.

show abstract

Robust Soft Semi-supervised Discriminant Projection for Feature Learning

Wang

Zhang

2016

Neural Information Processing

View full text Add to dashboard Cite

Semi-Supervised Image Classification by Nonnegative Sparse Neighborhood Propagation

Cited by 35 publications

References 26 publications

Learning Hybrid Representation by Robust Dictionary Learning in Factorized Compressed Space

Learning Hybrid Representation by Robust Dictionary Learning in Factorized Compressed Space

Robust Triple-Matrix-Recovery-Based Auto-Weighted Label Propagation for Classification

Robust Soft Semi-supervised Discriminant Projection for Feature Learning

Contact Info

Product

Resources

About