Surveillance face hallucination via variable selection and manifold learning

Jiang, Junjun; Hu, Ruimin; Han, Zhen; Lü, Tao; Huang, Kebin

doi:10.1109/iscas.2012.6271859

Cited by 6 publications

(7 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Manifold Manifold based methods [151] (2004), [206], [207], [310], [311], [327], [329], [343], [382], [386], [403], [404], [469], [495], [521], [553], [561], [568], [569], [570] assume that the HR and LR images form manifolds with similar local geometries in two distinct feature spaces [343]. Similar to PCA, these methods are also usually used for dimensionality reduction.…”

Section: Learning Based Single Image Sr Algorithmsmentioning

confidence: 99%

“…In the first step, they are combined with a MAP method [310], [311], [495], [568], [569], [570] or a Markov based learning method [206] like those in [65], [66], [76], [102], [146], [203] to apply a global constraint over the super-resolved image. In the second step, they use a different technique like Kernel Ridge Regression (KRR) [343], [495], graph embedding [568], radial basis function and partial least squares (RBF-PLS) regression [569], [570] to apply local constraints to the super-resolved image by finding the transformation between low and HR residual patches.…”

Section: Learning Based Single Image Sr Algorithmsmentioning

confidence: 99%

See 1 more Smart Citation

Super-resolution: a comprehensive survey

Nasrollahi

Moeslund

2014

Machine Vision and Applications

600

292

View full text Add to dashboard Cite

Super-resolution, the process of obtaining one or more high-resolution images from one or more low-resolution observations, has been a very attractive research topic over the last two decades. It has found practical applications in many real world problems in different fields, from satellite and aerial imaging to medical image processing, to facial image analysis, text image analysis, sign and number plates reading, and biometrics recognition, to name a few. This has resulted in many research papers, each developing a new superresolution algorithm for a specific purpose. The current comprehensive survey provides an overview of most of these published works by grouping them in a broad taxonomy. For each of the groups in the taxonomy, the basic concepts of the algorithms are first explained and then the paths through which each of these groups have evolved are given in detail, by mentioning the contributions of different authors to the basic concepts of each group. Furthermore, common issues in super-resolution algorithms, such as imaging models and registration algorithms, optimization of the cost functions employed, dealing with color information, improvement factors, assessment of super-resolution algorithms, and the most commonly employed databases are discussed.

show abstract

Section: Learning Based Single Image Sr Algorithmsmentioning

confidence: 99%

Section: Learning Based Single Image Sr Algorithmsmentioning

confidence: 99%

Super-resolution: a comprehensive survey

Nasrollahi

Moeslund

2014

Machine Vision and Applications

600

292

View full text Add to dashboard Cite

show abstract

“…These methods solve the SR problem by learning the relationship between HR and LR image patches or coefficients from a training set containing HR and LR image pairs. The target HR image can be inferred explicitly [10][11][12][13] or implicitly, which represent the input LR patch sparsely [13], collaboratively [14], and locally [15][16][17]. In [18], face image is considered as a whole.…”

Section: Introductionmentioning

confidence: 99%

Rigid Regression for Facial Image Interpolation with Local Structure Prior

Shao

et al. 2014

2014 Sixth International Conference on Intelligent Human-Machine Systems and Cybernetics

View full text Add to dashboard Cite

As an important biometric technique, face recognition has received substantial attention from both research communities and the market in the past three decades. However, the poor resolution of captured face images can seriously degrade the performance of conventional FR systems. To this end, many image super-resolution (SR) methods have been proposed recently. In this paper, we regard the face image SR as a kind of image interpolation problem for domain specific image. A missing intensity interpolation method based on Rigid Regression (RR) using a Local Structure Prior (LSP) is presented, which is called RRLSP for short. In order to interpolate missing intensities within a target high-resolution (HR) image, we assume that face image patches at the same position share the similar local structure, which is called LSP, and use the RR to learn the relationship between the low-resolution (LR) pixels and the missing HR pixels of one position patch with the LSP. Performance comparison with existing algorithms shows the effectiveness of the proposed method for face image SR in general.

show abstract

“…Super-resolution algorithms can be divided into two categories: reconstruction based algorithms and learning based algorithms. Algorithms based on learning gain more attention for its higher magnification and better performance [1][2][3][5][6][7][8][9][10][11][12][13]. In these learning based methods, they can reconstruct a HR image from an input image with priors from HR and LR face image training set.…”

Section: Introductionmentioning

confidence: 99%

“…However, the assumption must meet two premises: first, the sample data is dense sampling from the manifold space; second, the sample data is noiseless. This paper only focuses on the former; please refer to our previous work [12][13], which deal with the noisy manifold learning problem. Under the condition that training sample size is small, sample data can just constitute a sparse space of the highdimensional face manifold space, and even the most adjacent point can hardly be treated as local.…”

Section: Introductionmentioning

confidence: 99%

Face image super-resolution via nearest feature line

Han

Jiang

et al. 2012

Proceedings of the 20th ACM International Conference on Multimedia

Self Cite

View full text Add to dashboard Cite

In this paper, we propose a manifold learning based algorithm using 'Nearest Feature Line -NFL' to hallucinate high-resolution face image. According to the fact that existing NFL can effectively characterize the geometrical proportions to the face samples, we propose using NFL metric to define the neighborhood relations between face samples. Our algorithm can solve the problem that traditional method cannot effectively reveal the similar local geometry between high-resolution and low-resolution face manifolds under the condition that the training sample size is small. Moreover, in order to enhance the representation capacity of available face samples and reduce the computational complexity, we select neighborhood samples for each input LR image. Experimental results demonstrate that our algorithm can generates clearer local feature details, and the PSNR is 1.4 dB higher than that of the best manifold learning based method reported so far.

show abstract

Surveillance face hallucination via variable selection and manifold learning

Cited by 6 publications

References 8 publications

Super-resolution: a comprehensive survey

Super-resolution: a comprehensive survey

Rigid Regression for Facial Image Interpolation with Local Structure Prior

Face image super-resolution via nearest feature line

Contact Info

Product

Resources

About