Chunhong Pan

Chinese Academy of Sciences, Peping, Beijing, China

Are you Chunhong Pan?

Claim your profile

Publications (104)109.64 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: Hyperspectral unmixing (HU) plays a fundamental role in a wide range of hyperspectral applications. However, HU is still highly challenging due to the large solution space and the common presence of outlier channels. In this work, we propose a novel model by emphasizing both robust representation and learning-based sparsity. To relieve the side effects of badly degraded channels, a $\ell_{2,1}$-norm based robust loss is employed. Besides, the learning-based sparsity method is exploited by simultaneously learning the HU results and a sparse guidance map. Through this guidance map, the individually mixed level of each pixel is described and respected by imposing an adaptive sparsity constraint according to the mixed level of each pixel. Compared with the state-of-the-art method, such implementation is better suited to the real situation, thus expected to achieve better HU performances. The resulted objective is highly non-convex and non-smooth, and so it is hard to deal with. As a profound theoretical contribution, we propose an efficient algorithm to solve it. Meanwhile, the convergence proofs and the computational complexity analysis are systematically and theoretically provided. Extensive evaluations demonstrate that our method is highly promising for the HU task---it achieves very accurate guidance maps and extraordinarily better HU performances compared with the state-of-the-art methods.
    09/2014;
  • [Show abstract] [Hide abstract]
    ABSTRACT: Feature description for local image patch is widely used in computer vision. While the conventional way to design local descriptor is based on expert experience and knowledge, learning based methods for designing local descriptor become more and more popular because of their good performance and data-driven property. This paper proposes a novel data-driven method for designing binary feature descriptor, which we call Receptive Fields Descriptor (RFD). Technically, RFD is constructed by thresholding responses of a set of receptive fields, which are selected from a large number of candidates according to their distinctiveness and correlations in a greedy way. By using two different kinds of receptive fields (namely Rectangular pooling area and Gaussian pooling area) for selection, we obtain two binary descriptors RFDR and RFDG accordingly. Image matching experiments on the well known Patch Dataset [1] and Oxford Dataset [2] demonstrate that RFD significantly outperforms the state-of-the-art binary descriptors, and is comparable to the best float-valued descriptors at a fraction of processing time. Finally, experiments on object recognition tasks confirm that both RFDR and RFDG successfully bridge the performance gap between binary descriptors and their floating-point competitors.
    IEEE Transactions on Image Processing 04/2014; · 3.20 Impact Factor
  • Lingfeng Wang, Chunhong Pan
    [Show abstract] [Hide abstract]
    ABSTRACT: Nonrigid image registration plays an important role in medical imaging systems. In complex circumstances, it is still a challenging task, especially when input images are corrupted by the spatially-varying intensity distortion. To address this difficulty, we propose a novel locally linear reconstruction based dissimilarity measurement (LLR). The core idea behind the LLR is that in each local region, the aligned floating image is linearly reconstructed by the reference image. Due to LLR, we present a multi-resolution nonrigid image registration algorithm. Extensive experiments on various simulated and real medical images have demonstrated that our algorithm outperforms the state-of-the-art approaches in registration accuracy.
    Neurocomputing 01/2014; 145:303–315. · 1.63 Impact Factor
  • Lingfeng Wang, Chunhong Pan
    [Show abstract] [Hide abstract]
    ABSTRACT: It is still a challenging task to segment real-world images, since they are often distorted by unknown noise and intensity inhomogeneity. To address these problems, we propose a novel segmentation algorithm via a local correntropy-based K-means (LCK) clustering. Due to the correntropy criterion, the clustering algorithm can decrease the weights of the samples that are away from their clusters. As a result, LCK based clustering algorithm can be robust to the outliers. The proposed LCK clustering algorithm is incorporated into the region-based level set segmentation framework. The iteratively re-weighted algorithm is used to solve the LCK based level set segmentation method. Extensive experiments on synthetic and real images are provided to evaluate our method, showing significant improvements on both noise sensitivity and segmentation accuracy, as compared with the state-of-the-art approaches.
    Pattern Recognition. 01/2014; 47(5):1917–1925.
  • [Show abstract] [Hide abstract]
    ABSTRACT: Face recognition has been popular in the pattern recognition field for decades, but it is still a difficult problem due to the various image distortions. Recently, sparse representation based classification (SRC) was proposed as a novel image classification approach, which is very effective with sufficient training samples for each class. However, the performance drops when the number of training samples is limited. In this paper, we show that effective local image features and appropriate nonlinear kernels are needed in deriving a better classification method based on sparse representation. Thus, we propose a novel kernel SRC framework and utilize effective local image features in this framework for robust face recognition. First, we present a kernel coordinate descent (KCD) algorithm for the LASSO problem in the kernel space, and we successfully integrate it in the SRC framework (called KCD-SRC) for face recognition. Second, we employ local image features and develop both pixel-level and region-level kernels for KCD-SRC based face recognition, making it discriminative and robust against illumination variations and occlusions. Extensive experiments are conducted on three public face databases (Extended YaleB, CMU-PIE and AR) under illumination variations, noise corruptions, continuous occlusions, and registration errors, demonstrating excellent performances of the KCD-SRC algorithm combining with the proposed kernels.
    Neurocomputing 01/2014; 133:141–152. · 1.63 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Hyperspectral Unmixing (HU) has received increasing attention in the past decades due to its ability of unveiling information latent in hyperspectral data. Unfortunately, most existing methods fail to take advantage of the spatial information in data. To overcome this limitation, we propose a Structured Sparse regularized Nonnegative Matrix Factorization (SS-NMF) method based on the following two aspects. First, we incorporate a graph Laplacian to encode the manifold structures embedded in the hyperspectral data space. In this way, the highly similar neighboring pixels can be grouped together. Second, the lasso penalty is employed in SS-NMF for the fact that pixels in the same manifold structure are sparsely mixed by a common set of relevant bases. These two factors act as a new structured sparse constraint. With this constraint, our method can learn a compact space, where highly similar pixels are grouped to share correlated sparse representations. Experiments on real hyperspectral data sets with different noise levels demonstrate that our method outperforms the state-of-the-art methods significantly.
    ISPRS Journal of Photogrammetry and Remote Sensing 01/2014; 88:101–118. · 3.31 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a novel region-based framework for multifocus image fusion. The core idea is to segment the in-focus regions from the input images and merge them up to produce an all-in-focus image. To this end, we propose three intuitive constraints on the fusion process and model them into three energy terms, i.e., reconstruction error, out-of-focus energy and smoothness regularization. The three terms are then formulated into an optimization framework problem to solve a segmentation map. We also propose a greedy algorithm to minimize the objective function, which alternatively updates each pixel in the segmentation map using a coarse-to-fine strategy. The fused image is finally generated by combining the segmented in-focus regions in each source image via the segmentation map. Our approach can yield a seamless result with much fewer ringing artifacts. Comparative experiments based on various synthesized and real images demonstrate that our approach outperforms the state-of-the-art methods.
    Neurocomputing. 01/2014; 140:193–209.
  • Source
    Shaoguo Liu, Haibo Wang, Jue Wang, Chunhong Pan
    The Visual Computer 01/2014; · 0.91 Impact Factor
  • Source
    Shaoguo Liu, Haibo Wang, Yiyi Wei, Chunhong Pan
    [Show abstract] [Hide abstract]
    ABSTRACT: Homography estimation is a fundamental problem in the field of computer vision. For estimating the homography between two images, one of the key issues is to match keypoints in the reference image to the keypoints in the moving image. In order to match keypoints in real time, binary image descriptor, due to its low matching and storage costs, emerges as a more and more popular tool. On achieving the low costs, binary descriptor sacrifices the discriminative power of using floating points. In this paper, we present BB-Homography, a new approach that fuses fast Binary descriptor matching and Bipartite graph for Homography estimation. Starting with binary descriptor matching, BB-Homography uses bipartite graph matching algorithm to refine the matching results, which are finally passed over to estimate homography. On realizing the correlation between keypoint correspondence and homography estimation, BB-Homography iteratively performs the graph matching and the homography estimation such that they can refine each other at each iteration. In particular, based on spectral graph, a fast bipartite graph matching algorithm is developed for lowering the time cost of BB-Homography. BB-Homography is extensively evaluated on both public benchmarks and live-captured video streams, which consistently shows that BB-Homography outperforms conventional methods for homography estimation.
    IEEE Transactions on Circuits and Systems for Video Technology 01/2014; · 1.82 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Images captured in foggy weather conditions often suffer from bad visibility. In this paper, we propose an efficient regularization method to remove hazes from a single input image. Our method benefits much from an exploration on the inherent boundary constraint on the transmission function. This constraint, combined with a weighted L_1-norm based contextual regularization, is modeled into an optimization problem to estimate the unknown scene transmission. A quite efficient algorithm based on variable splitting is also presented to solve the problem. The proposed method requires only a few general assumptions and can restore a high-quality haze-free image with faithful colors and fine image details. Experimental results on a variety of haze images demonstrate the effectiveness and efficiency of the proposed method.
    Proceedings of the 2013 IEEE International Conference on Computer Vision; 12/2013
  • Lingfeng Wang, Chunhong Pan
    [Show abstract] [Hide abstract]
    ABSTRACT: Magnetic resonance (MR) image segmentation is a crucial step in surgical and treatment planning. In this paper, we propose a level-set-based segmentation method for MR images with intensity inhomogeneous problem. To tackle the initialization sensitivity problem, we propose a new image-guided regularization to restrict the level set function. The maximum a posteriori inference is adopted to unify segmentation and bias field correction within a single framework. Under this framework, both the contour prior and the bias field prior are fully used. As a result, the image intensity inhomogeneity can be well solved. Extensive experiments are provided to evaluate the proposed method, showing significant improvements in both segmentation and bias field correction accuracies as compared with other state-of-the-art approaches.
    Magnetic Resonance Imaging 11/2013; · 2.06 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Change detection of VHR (Very High Resolution) images is very difficult due to the impacts caused by the seasonal changes, the imaging condition, and so on. To address the above difficulty, a novel unsupervised change detection algorithm is proposed based on deep learning, where the complex correspondence between the images is established by Auto-encoder Model. By taking advantages of the powerful ability of deep learning in compensating the impacts implicitly, the multi-temporal images can be compared fairly. Experiments demonstrate the effectiveness of the proposed approach.
    Proc SPIE 10/2013;
  • Bin Fan, Chunlei Huo, Chunhong Pan, Qingqun Kong
    [Show abstract] [Hide abstract]
    ABSTRACT: Although feature-based methods have been successfully developed in the past decades for the registration of optical images, the registration of optical and synthetic aperture radar (SAR) images is still a challenging problem in remote sensing. In this letter, an improved version of the scale-invariant feature transform is first proposed to obtain initial matching features from optical and SAR images. Then, the initial matching features are refined by exploring their spatial relationship. The refined feature matches are finally used for estimating registration parameters. Experimental results have shown the effectiveness of the proposed method.
    IEEE Geoscience and Remote Sensing Letters 07/2013; 10(4):657-661. · 1.82 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Hyperspectral unmixing is one of the crucial steps for many hyperspectral applications. The problem of hyperspectral unmixing has proven to be a difficult task in unsupervised work settings where the endmembers and abundances are both unknown. What is more, this task becomes more challenging in the case that the spectral bands are degraded with noise. This paper presents a robust model for unsupervised hyperspectral unmixing. Specifically, our model is developed with the correntropy based metric where the non-negative constraints on both endmembers and abundances are imposed to keep physical significance. In addition, a sparsity prior is explicitly formulated to constrain the distribution of the abundances of each endmember. To solve our model, a half-quadratic optimization technique is developed to convert the original complex optimization problem into an iteratively re-weighted NMF with sparsity constraints. As a result, the optimization of our model can adaptively assign small weights to noisy bands and give more emphasis on noise-free bands. In addition, with sparsity constraints, our model can naturally generate sparse abundances. Experiments on synthetic and real data demonstrate the effectiveness of our model in comparison to the related state-of-the-art unmixing models.
    05/2013;
  • Lingfeng Wang, Huaiyu Wu, Chunhong Pan
    [Show abstract] [Hide abstract]
    ABSTRACT: Intensity inhomogeneity often causes considerable difficulties in image segmentation. To tackle this problem, we propose a new region-based level set method. The proposed method considers the local image information by describing it as a novel local signed difference (LSD) energy, which possesses both local separability and global consistency. The LSD energy term is integrated into an objective energy functional, which is minimized via a level set evolution process. Extensive experiments are performed to evaluate the proposed method, showing improvements in both accuracy and efficiency, as compared with the state-of-the-art approaches.
    Pattern Recognition Letters 04/2013; 34(6):637–645. · 1.27 Impact Factor
  • Jun Bai, Shiming Xiang, Chunhong Pan
    [Show abstract] [Hide abstract]
    ABSTRACT: The goal of this paper is to apply graph cut (GC) theory to the classification of hyperspectral remote sensing images. The task is formulated as a labeling problem on Markov random field (MRF) constructed on the image grid, and GC algorithm is employed to solve this task. In general, a large number of user interactive strikes are necessary to obtain satisfactory segmentation results. Due to the spatial variability of spectral signatures, however, hyperspectral remote sensing images often contain many tiny regions. Labeling all these tiny regions usually needs expensive human labor. To overcome this difficulty, a pixelwise fuzzy classification based on support vector machine (SVM) is first applied. As a result, only pixels with high probabilities are preserved as labeled ones. This generates a pseudouser strike map. This map is then employed for GC to evaluate the truthful likelihoods of class labels and propagate them to the MRF. To evaluate the robustness of our method, we have tested our method on both large and small training sets. Additionally, comparisons are made between the results of SVM, SVM with stacking neighboring vectors, SVM with morphological preprocessing, extraction and classification of homogeneous objects, and our method. Comparative experimental results demonstrate the validity of our method.
    IEEE Transactions on Geoscience and Remote Sensing 02/2013; 51(2):803-817. · 3.47 Impact Factor
  • Lingfeng Wang, Huaiyu Wu, Chunhong Pan
    [Show abstract] [Hide abstract]
    ABSTRACT: It is still a very challenging task to recognize a face in a real world scenario, since the face may be corrupted by many unknown factors. Among them, illumination variation is an important one, which will be mainly discussed in this paper. First, the illumination variations caused by shadow or overexposure are regarded as a multiplicative scaling image over the original face image. The purpose of introducing scaling vector (or scaling image) is to enhance the pixels in shadow regions, while depress the pixels in overexposure regions. Then, based on the scaling vector, we propose a novel tone-aware sparse representation (TASR) model. Finally, a EM-like algorithm is proposed to solve the proposed TASR model. Extensive experiments on the benchmark face databases show that our method is more robust against illumination variations.
    Automatic Face and Gesture Recognition (FG), 2013 10th IEEE International Conference and Workshops on; 01/2013
  • Kun Ding, Chunlei Huo, Yuan Xu, Chunhong Pan
    [Show abstract] [Hide abstract]
    ABSTRACT: The difficulty of Very High Resolution (VHR) image change detection is mainly due to the low separability between the changed and unchanged class. The traditional approaches usually address the problem by solving the feature extraction and classification separately, which cannot ensure that the classification algorithm makes the best use of the features. Considering this, we propose a novel approach that combines the feature extraction and the classification task by utilizing the sparse representation algorithm with discriminative dictionary. Experiments on real data sets show that our method achieves effective results.
    Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on; 01/2013 · 4.63 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Local image descriptors are one of the key components in many computer vision applications. Recently, binary descriptors have received increasing interest of the community for its efficiency and low memory cost. The similarity of binary descriptors is measured by Hamming distance which has equal emphasis on all elements of binary descriptors. This paper improves the performance of binary descriptors by learning a weighted Hamming distance for binary descriptors with larger weights assigned to more discriminative elements. What is more, the weighted Hamming distance can be computed as fast as the Hamming distance on the basis of a pre-computed look-up-table. Therefore, the proposed method improves the matching performance of binary descriptors without sacrificing matching speed. Experimental results on two popular binary descriptors (BRIEF [1] and FREAK [2]) validate the effectiveness of the proposed method.
    Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on; 01/2013 · 4.63 Impact Factor
  • Source
    Shaoguo Liu, Ying Wang, Haibo Wang, Chunhong Pan
    [Show abstract] [Hide abstract]
    ABSTRACT: Depth maps provided by Microsoft Kinect always suffer from large dark holes around depth boundaries and occasional missing pixels in non-occluded regions, as well as noise, which prevent their further usage in real-world applications. In this paper, we propose a graph Laplacian based framework to restore the missing pixels according to the strong correlation between color image and depth map. To preserve sharp edges and remove noise, the T V21 (Total Variation) prior of depth maps is then introduced to the framework. Finally, an efficient and effective iterative optimization method with a closed-form solution at each iteration is presented to address this issue. Experiments conducted on both real scene images and synthetic images demonstrate that our approach gives better performance than commonly-used depth inpainting schemes.
    ACPR; 01/2013