Kebin Huang

Position-Patch Based Face Hallucination via Locality-Constrained Representation

Instead of using probabilistic graph based or manifold learning based models, some approaches based on position-patch have been proposed for face hallucination recently. In order to obtain the optimal weights for face hallucination, they represent image patches through those patches at the same position of training face images by employing least square estimation or convex optimization. However, they can hope neither to provide unbiased solutions nor to satisfy locality conditions, thus the obtained patch representation is not the best. In this paper, a simpler but more effective representation scheme- Locality-constrained Representation (LcR) has been developed, compared with the Least Square Representation (LSR) and Sparse Representation (SR). It imposes a locality constraint onto the least square inversion problem to reach sparsity and locality simultaneously. Experimental results demonstrate the superiority of the proposed method over some state-of-the-art face hallucination approaches.

Efficient Single Image Super-Resolution via Graph Embedding

We explore in this paper efficient algorithmic solutions to single image super-resolution (SR). We propose the GESR, namely Graph Embedding Super-Resolution, to super-resolve a high-resolution (HR) image from a single low-resolution (LR) observation. The basic idea of GESR is to learn a projection matrix mapping the LR image patch to the HR image patch space while preserving the intrinsic geometrical structure of original HR image patch manifold. While GESR resembles other manifold learning-based SR methods in persevering the local geometric structure of HR and LR image patch manifold, the innovation of GESR lies in that it preserves the intrinsic geometrical structure of original HR image patch manifold rather than LR image patch manifold, which may be contaminated because of image degeneration (e.g., blurring, down-sampling and noise). Experiments on benchmark test images show that GESR can achieve very competitive performance as Neighbor Embedding based SR (NESR) and Sparse representation based SR (SSR). Beyond subjective and objective evaluation, all experiments show that GESR is much faster than both NESR and SSR.

Noise robust position-patch based face super-resolution via Tikhonov regularized neighbor representation

A novel position-patch face super-resolution method for face super-resolution is developed.We developed a novel patch representation method based on Tikhonov regularized neighbor representation.The proposed patch representation model achieves sparsity and locality simultaneously.The effectiveness of our algorithms was shown on public face databases and some real-world images. In human-machine interaction, human face is one of the core factors. However, due to the limitations of imaging conditions and low-cost imaging sensors, the captured faces are often low-resolution (LR). This will seriously degrade the performance of face detection, expression analysis, and face recognition, which are the basic problems in human-machine interaction applications. Face super-resolution (SR) is the technology of inducing a high-resolution (HR) face from the observed LR one. It has been a hot topic of wide concern recently. In this paper, we present a novel face SR method based on Tikhonov regularized neighbor representation (TRNR). It can overcome the technological bottlenecks (e.g., instable solutionand noise sensitive) of the patch representationscheme in traditional neighbor embedding based image SR methods. Specifically, we introduce the Tikhonov regularization term to regularize the representation of the observation LR patches, leading to a unique and stable solution for the least squares problem. Furthermore, we show a connection of the proposed model to the neighbor embedding model, least squares representation,sparse representation, and locality-constrained representation. Extensive experiments on face SR are carried out to validate the generality, effectiveness, and robustness of the proposed algorithm. Experimental results on the public FEI face database and surveillance images show that the proposed method achieves better performance in terms of reconstruction error and visual quality than existing state-of-the-art methods.

Graph discriminant analysis on multi-manifold (GDAMM): A novel super-resolution method for face recognition

How to efficiently recognize low-resolution (LR) probe images of one face recognition system, in which high-resolution (HR) gallery of faces is enrolled, is still an open problem. In this paper, we develop a novel super-resolution method, namely Graph Discriminant Analysis on Multi-Manifold (GDAMM), to super-resolved the HR version of a LR probe image and then perform matching at the resolution of the HR gallery. Unlike classical super-resolution approaches considering only the data fidelity, GDAMM takes the advantages of both manifold learning and discriminant analysis to integrate the data constraint and discriminant constraint, seeking the mapping between LR images and HR ones. In the reconstructed HR image space, faces of one person in the same manifold are close and those in different manifolds are far apart. Experiments on Extended Yale-B database and AR face database demonstrate that the learned discriminant information is essential for improving recognition accuracy. Through the contrastive experiment, the results (recognition rates) indicate that the proposed GDAMM method can greatly surpass classical super-resolution approaches, even outperforming the ideal case of having probe images of HR gallery by a big margin (nearly 9% on Extended Yale-B database and 8% on AR face database).

A super-resolution method for low-quality face image through RBF-PLS regression and neighbor embedding

In this paper, a new two-step method is proposed to infer a high-quality and high-resolution (HR) face image from a low-quality and low-resolution (LR) observation based on training samples in the database. First, a global face image is reconstructed based on the non-linear relationship between LR and HR face images, which is established according to radial basis function and partial least squares (RBF-PLS) regression. Based on the reconstructed global face patches manifold (formed by the image patches at the same position of all global face images), whose local geometry is more consistent with that of original HR face patches manifold than noisy LR one is, the Neighbor Embedding is applied to induce the target HR face image by preserving the similar local geometry between global face patches manifold and the original HR face patches manifold. A comparison of some state-of-the-art methods shows the superiority of our method, and experiments also demonstrate the effectiveness both under simulation and real conditions.

Surveillance face hallucination via variable selection and manifold learning

In this paper, we propose a new two-step face hallucination method to induce a high-resolution (HR) face image from a low-resolution (LR) observation. Especially for low-quality surveillance face image, an RBF-PLS based variable selection method is presented for the reconstruction of global face image. Further more, in order to compensate for the reconstruction errors, which are lost high frequency detailed face features, the Neighbor Embedding (NE) based residue face hallucination algorithm is used. Compared with current methods, the proposed RBF-PLS based method can generate a global face more similar to the original face and less sensitive to noise, moreover, the NE algorithm can reduce the reconstruction errors caused by misalignment on the basis of a carefully designed search strategy. Experiments show the superiority of the proposed method compared with some state-of-the-art approaches and the efficacy both in simulation and real surveillance condition.

A face super-resolution method based on illumination invariant feature

Human faces in surveillance video images usually have low resolution and poor quality. They need to be reconstructed in super-resolution for identification. The traditional subspace-based face super-resolution algorithms are sensitive to light. For solving the problem, this paper proposes a face super-resolution method based on illumination invariant feature. The method firstly extracts the illumination invariant features of an input low resolution image by using adaptive L1–L2 total variation model and self-quotient image in logarithmic domain. Then it projects the feature onto non-negative basis obtained by Nonnegative Matrix Factorization(NMF) in face image database. Finally it reconstructs the high resolution face images under the framework of Maximum A Posteriori (MAP) probability. Experimental results demonstrate that the proposed method outperforms the compared methods both in subjective and objective quality under poor light conditions.

HRM graph constrained dictionary learning for face image super-resolution

Sparse coding based face image Super-Resolution (SR) approaches have received increasing amount of interest recently. However, most of the existing sparse coding based approaches fail to consider the geometrical structure of the face space, as a result, artificial effects on reconstructed High Resolution (HR) face images come into being. In this paper, a novel sparse coding based face image SR method is proposed to reconstruct a HR face image from a Low Resolution (LR) observation. In training stage, it aims to get a more expressive HR-LR dictionary pair for certain input LR patch. The intrinsic geometric structure of training samples is incorporated into the sparse coding procedure for dictionary learning. Unlike the existing SR methods which use the graph constructed in LR Manifold (LRM) as regularization term, the proposed method uses graph constructed in HR Manifold (HRM) as regularization term. In reconstruction stage, K selection mean constrains is used in l1 convex optimization, aiming at finding an optimal weight for HR face image patch reconstruction. Experimental results on both simulation and real world images suggest that our proposed one achieves better quality when compared with other state-of-the-art methods.