Learning dictionary on manifolds for image classification

Liu, Baodi; Wang, Yu-Xiong; Zhang, Yujin; Shen, Bin

doi:10.1016/j.patcog.2012.11.018

Cited by 69 publications

(19 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, it is helpful for SR that similar samples constitutionally exist in the same subspace while heterogeneous samples potentially lie in different subspaces. In high-dimensional case, the overcomplete features spontaneously result in the small overlap between t S and other subspaces even the disjoined subspaces and thus promote SR to deliver the believable neighbor selection of t, which is supported by the studies of [35,55]. However, as the difference value m d − decreases, the overlap between t S and other subspaces increases and thereby gives rise to the unreliability of neighbor selection by SR.…”

Section: Analysis Of Srmentioning

confidence: 90%

Linear dimensionality reduction based on Hybrid structure preserving projections

2016

View full text Add to dashboard Cite

Section: Analysis Of Srmentioning

confidence: 90%

Linear dimensionality reduction based on Hybrid structure preserving projections

2016

View full text Add to dashboard Cite

“…Single scale SIFT ScSPM(1024) [40,6] 82.74 ± 1.46 80.28 ± 0.93 EMK [1] 74.56 ± 1.32 NA KSR [4] 84.92 ± 0.78 83.68 ± 0.61 SCSR(1024) [17] 87.97 ± 1.11 81.51 ± 0.32 DLSM(1024) [18] 86.82 ± 1.04 83.40 ± 0.44 DLMM(1024) [18] 86 we call multiple scales SIFT. 128 dimensional SIFT descriptors are obtained and normalized to 1 with 2 -norm.…”

Section: Uiuc-sports Scene 15mentioning

confidence: 99%

“…One example task is image classification [18,19,40], which aims to associate images with semantic labels automatically. The most common framework is the discriminative model [12,38,40].…”

Section: Introductionmentioning

confidence: 99%

Self-explanatory Sparse Representation for Image Classification

Liu

Wang

Shen

et al. 2014

Computer Vision – ECCV 2014

Self Cite

View full text Add to dashboard Cite

Abstract. Traditional sparse representation algorithms usually operate in a single Euclidean space. This paper leverages a self-explanatory reformulation of sparse representation, i.e., linking the learned dictionary atoms with the original feature spaces explicitly, to extend simultaneous dictionary learning and sparse coding into reproducing kernel Hilbert spaces (RKHS). The resulting single-view self-explanatory sparse representation (SSSR) is applicable to an arbitrary kernel space and has the nice property that the derivatives with respect to parameters of the coding are independent of the chosen kernel. With SSSR, multiple-view self-explanatory sparse representation (MSSR) is proposed to capture and combine various salient regions and structures from different kernel spaces. This is equivalent to learning a nonlinear structured dictionary, whose complexity is reduced by learning a set of smaller dictionary blocks via SSSR. SSSR and MSSR are then incorporated into a spatial pyramid matching framework and developed for image classification. Extensive experimental results on four benchmark datasets, including UIUC-Sports, Scene 15, Caltech-101, and Caltech-256, demonstrate the effectiveness of our proposed algorithm.

show abstract

“…Table 7 shows the results, on the Caltech256 dataset, of the 1 BRD, the competing histogram distances, the logistic regression, the method in [12], the method based on the dictionary learning on single manifold (DLSM) in [47], the method based on the dictionary learning on multiple manifolds (DLMM) in [47], and the method in [11]. The experimental setups for the different histogram distances are exactly the same to avoid bias.…”

Section: Global Synthesismentioning

confidence: 99%

Bin Ratio-Based Histogram Distances and Their Application to Image Classification

Xie

et al. 2014

IEEE Trans. Pattern Anal. Mach. Intell.

View full text Add to dashboard Cite

Abstract:Large variations in image background may cause partial matching and normalization problems for histogram-based representations, i.e. the histograms of the same category may have bins which are significantly different, and normalization may produce large changes in the differences between corresponding bins. In this paper, we deal with this problem by using the ratios between bin values of histograms, rather than bin values' differences which are used in the traditional histogram distances. We propose a bin ratio-based histogram distance (BRD), which is an intra-cross-bin distance, in contrast with previous bin-to-bin distances and cross-bin distances. The BRD is robust to partial matching and histogram normalization, and captures correlations between bins with only a linear computational complexity. We combine the BRD with the 1 histogram distance and the

show abstract

Learning dictionary on manifolds for image classification

Cited by 69 publications

References 19 publications

Linear dimensionality reduction based on Hybrid structure preserving projections

Linear dimensionality reduction based on Hybrid structure preserving projections

Self-explanatory Sparse Representation for Image Classification

Bin Ratio-Based Histogram Distances and Their Application to Image Classification

Contact Info

Product

Resources

About