Lian Qiusheng

Yan Shan University | YSU · School of Information Science and Engineering

In low-light conditions, captured images often suffer from poor visibility and visual perception. Convolutional neural networks (CNNs) have shown promising performance in enhancing low-light images. However, CNNs encounter difficulties in modeling global information, which is crucial for effective enhancement. This paper proposes an end-to-end mult...

Low-light image enhancement algorithms have been widely developed. Nevertheless, using long exposure under low-light conditions will lead to motion blurs of the captured images, which presents a challenge to address low-light enhancement and deblurring jointly. A recent effort called LEDNet addresses these issues by designing a encoder-decoder pipe...

Reference-free low-light image enhancement methods only employ low-light images during training, thereby significantly alleviating the over-reliance on obtaining paired or unpaired datasets. Existing reference-free low-light image enhancement approaches still struggle to strike a balance between enhancing vivid color and suppressing noise in low-li...

Sparse-view computed tomography (SVCT) is regarded as a promising technique to accelerate data acquisition and reduce radiation dose. However, in the presence of metallic implants, SVCT inevitably makes the reconstructed CT images suffer from severe metal artifacts and streaking artifacts due to the lack of sufficient projection data. Previous stan...

Computed tomography (CT) images are often corrupted by undesirable artifacts due to the presence of metallic implants, creating the problem of metal artifact reduction (MAR). Existing deep learning-based efforts of tackling this problem share two main common limitations, limiting their practical applications. Firstly, single domain knowledge is ins...

Quanta image sensors (QIS) are a new type of single-photon imaging device that can oversample the light field to generate binary bit-streams. The reconstruction for QIS refers to the recovery of original scenes from these binary measurements. Conventional reconstruction algorithms for QIS generally depend solely on one instantiated prior and are ce...

Recent deep learning-based methods have achieved promising performance for computed tomography metal artifact reduction (CTMAR). However, most of them suffer from two limitations: (i) the domain knowledge is not fully embedded into the network training; (ii) metal artifacts lack effective representation models. The aforementioned limitations leave...

Probing the issue of phase retrieval has attracted researchers for many years, due to its wide range of application. Phase retrieval aims to recover an unknown signal from phase-free measurements. Classical alternative projection algorithms have the significant advantages of simplicity and few fine-tuning parameters. However, they suffer from non-c...

Deep convolutional neural networks (DCNN) have been widely used in the field of image denoising because of their fast inference and good performance. However, the design of networks for the DCNN is mostly empirical, and the interpretation and robustness of them remains a major challenge. Inspired by the total generalized variation method, this pape...

Block compressed sensing (BCS) is effective to process high-dimensional images or videos. Due to the block-wise sampling, most BCS methods only exploit local block priors and neglect inherent global image priors, thus resulting in blocky artifacts. To ameliorate this issue, this paper formulates a novel regularized optimization BCS model named BCS-...

Deep convolutional neural networks (CNNs) have been very successful in image denoising. However, with the growth of the depth of plain networks, CNNs may result in performance degradation. The lack of network depth leads to the limited ability of the network to extract image features and difficults to fuse the shallow image features into the deep i...

Data‐driven tight frames are popular for solving imaging inverse problems. However, the imaging quality is limited by the representation ability of single tight frame and thresholds tuned manually. In this work, a supervised dual tight frame learning framework fused with an elaborated deep thresholding network (DTN) is proposed, and the issue of lo...

Phase retrieval (PR), i.e., the recovery of the underlying image from the measurements without phase information, is a challenging task, especially at low signal to noise ratios (SNRs). Recent deep unrolling optimizations of tackling this task offer both computational efficiency and high-quality reconstructions. In this work, we involve a novel dee...

Compressed Sensing theory breaks through the limitation of the Nyquist sampling law and provides theoretical support for accelerating the imaging process of MRI while reconstructing high-quality images. It can use less sampling data through the reconstructed algorithm to restore the original signal. Reconstruction time and reconstruction algorithms...

Compressed sensing (CS) aims to precisely reconstruct the original signal from under-sampled measurements, which is a typical ill-posed problem. Solving such a problem is challenging and generally needs to incorporate suitable priors about the underlying signals. Traditionally, these priors are hand-crafted and the corresponding approaches generall...

Signal models play a paramount role in compressed sensing magnetic resonance imaging (CSMRI), which aims to accurately recover magnetic resonance (MR) images from highly undersampled measurements. In recent decade, lots of works exploit the sparsity and the low rank for CSMRI. However, there are some defects involving many finely-tuned parameters a...

Diffraction imaging problem, i.e. recovery of a high-resolution or high-quality image from the intensity diffraction pattern, arises in many science and engineering fields. Recent efforts to solve this problem are exploiting sparse representation techniques. However, existing sparse representation models cannot explore inherent priors of the images...

The image representation plays an important role in compressed sensing magnetic resonance imaging (CSMRI). However, how to adaptive sparsely represent images is a challenge for accurately reconstructing magnetic resonance (MR) images from highly undersampled data with noise. In order to further improve the reconstruction quality of the MR image, th...

The signal degradation due to the Poisson noise is a common problem in the low‐light imaging field. Recently, deep learning employing the convolution neural network for image denoising has drawn considerable attention owing to its favourable denoising performance. On the basis of the fact that the reconstruction of corrupted pixels can be facilitat...

The problem of recovering an image of interest from nonlinear measured data is challenging. To address this nonlinear imaging inverse problem, we propose a novel Plug-and-Play Regularization (PPR) approach that can exploit multiple priors. The underlying image and its filtered image by a denoiser should have similar structures. To exploit this simi...

We propose a purely automatic and accurate phase aberration compensation method in digital holographic microscopy for phase-contrast imaging of living cells. At first, a least-squares fitting procedure is implemented to correct a large proportion of the phase aberrations. Subsequently, the flat phase background is retrieved based on phase gradient...

Removing rain from a single image is a challenging task due to the absence of temporal information. Considering that a rainy image can be decomposed into the low-frequency ( LF ) and high-frequency ( HF ) components, where the coarse scale information is retained in the LF component and the rain streaks and texture correspond to the HF component, w...

How to explore many priors of the underlying image for diffraction imaging, i.e., recovering the image from recorded diffraction patterns, is an important issue. To address this issue, we present a sparse regularization model called SBM3D (sparse regularization model induced by block-matching and 3D filtering). The proposed SBM3D exploits the spars...

Simplicity and few finely tuned parameters are the main advantages of alternating projection (AP) methods, a fundamental class of phase retrieval (PR) methods in the optical imaging field. However, AP methods often suffer from low-quality imaging when few diffraction patterns are recorded. Regularized PR methods avoid this deficiency by using some...

Compressed-sensing magnetic resonance imaging (CSMRI) aims to reconstruct the magnetic resonance (MR) image from highly undersampled K-space data. In order to improve the reconstruction quality of the MR image, this paper proposes a new gradient-based tight frame (TFG) learning algorithm (TFG-MRI) for CSMRI. TFG-MRI effectively integrates the tight...

We propose a numerical and totally automatic phase aberration compensation method in digital holographic microscopy. The phase aberrations are extracted in a nonlinear optimization procedure in which the phase variation of the reconstructed object wave is minimized. Not only phase curvature but also high-order aberrations could be corrected without...

Coded diffraction patterns (CDPs) recorded by optical detectors are often affected by Poisson noise in optical applications. How to recover the image of interest from few noisy CDPs is a challenge. In this paper, a double sparse regularization (DSR) model that exploits both the gradient sparsity and the structured sparsity is proposed to recover th...

In this paper, a new method is proposed to identify solid oxide fuel cell using extreme learning machine–Hammerstein model (ELM–Hammerstein). The ELM–Hammerstein model consists of a static ELM neural network followed by a linear dynamic subsystem. First, the structure of ELM–Hammerstein model is determined by Lipschitz quotient criterion from input...

In this paper, an online identification method is proposed for nonlinear system identification based on extreme learning machine (ELM)–Hammerstein model. The ELM–Hammerstein model comprises an ELM neural network followed by a linear dynamic subsystem. This model is linear in parameters and nonlinear in the input. To speed up the convergence and mea...

Existing mixed near-field and far-field source localization algorithms rely on uniform linear array (ULA), and the maximum number of sources that they can detect is less than the sensor number. As a comparison, recently proposed nested array can provide higher number of consecutive lags with the same sensor number, enabling us to extend it to mixed...

Magnetic resonance imaging (MRI) has been widely employed in medical diagnosis since it enables superior visualization of anatomical structure with noninvasive and nonionizing radiation nature. However, during the data acquisition process of MRI, patients’ translational motion usually leads to phase changes of the observed data; moreover, the ampli...

A novel Synthetic aperture radar (SAR) signal processing technique has been proposed which refocused slow moving targets based on phase retrieval algorithm. After theoretical derivation, we can get that the raw data Fourier magnitude of slow moving targets is approximate to the stationary ones in the SAR system. By applying the Fourier magnitude of...

We consider the problem of two-dimensional (2-D) angles of arrival estimation using a newly proposed structure of nonuniform linear array, referred to as nested coprime array with compressed inter-element spacing (CACIS). By constructing a cross-correlation matrix of the received signals, the nested CACIS exhibits a larger number of degrees of free...

The problem of phase retrieval, namely, recovery of a signal only from the magnitude of its Fourier transform, or of any other linear transform. Due to the loss of phase information, this problem is ill-posed. Therefore, the prior knowledge is required to enable its accurate reconstruction. In this work, based on the framework of nonlinear compress...

In this paper, a novel two-dimensional (2-D) angle of arrival estimation algorithm based on sparse signal reconstruction is addressed with an L-sharped Array. By constructing a cross-correlation vector of the received signals, the 2-D angles estimation are obtained by -norm minimization and least square procedure, respectively. Without pair matchin...

Due to the influence of the platform random motion and electromagnetic propagation in turbulent media, the synthetic aperture radar (SAR) high resolution imaging for the sea scenes where there are large amounts of water returns with some target (land) returns is very difficult. To solve this problem, a SAR imaging method based on the improved phase...

The phase retrieval (PR) problem of recovering an image from its Fourier magnitudes is an important issue. Several PR algorithms have been proposed to address this problem. Recent efforts of exploiting sparsity were developed to improve the performance of PR algorithms, such as the reconstruction quality, robustness to noise, and convergence behavi...

At present, the sparse representation-based classification (SRC) methods of electroencephalograph (EEG) signal analysis have become an important approach for studying brain science. SRC methods mean that the target data is sparsely represented on the basis of a fixed dictionary or learned dictionary, and classified based on the reconstruction crite...

How to improve the reconstructed image quality using more prior knowledge of the image is still a crucial issue of compressed sensing. In this paper, we combine the synthesis sparse model and the cosparse analysis model proposed in recent years, and propose a novel reconstruction algorithm based on the sparsity of the image over a synthesis diction...

To tackle the problem that dark channel prior is invalid for large sky or bright objects regions of single hazy image and the problem that some hazes cannot be removed using guided filter, we propose a single image dehazing algorithm based on improved dark channel prior and guided filter. Firstly, we introduce a mixed dark channel based on dark cha...

Antilock braking system (ABS) has been designed to attain maximum negative acceleration and prevent the wheels from locking. Many efforts had been paid to design controller for ABS to improve the brake performance, especially when road condition changes. In this paper, an adaptive fuzzy fractional-order sliding mode controller (AFFOSMC) design meth...

We treat the phase retrieval (PR) problem of reconstructing the interest signal from its Fourier magnitude. Since the Fourier phase information is lost, the problem is ill-posed. Several techniques have been used to address this problem by utilizing various priors such as non-negative, support, and Fourier magnitude constraints. Recent methods expl...

Compressed sensing (CS) enables that magnetic resonance (MR) images can be exactly reconstructed from undersampled k-space data by exploiting the sparsity of MR images in some analytical sparsifying transform or some dictionary. Recent methods are exploiting adaptive patch-based dictionaries for image recovery by alternating between dictionary lear...

A novel compressed sensing algorithm based on learning analysis dictionary and optimizing measurement matrix from subspaces is proposed in this paper. The whole image space is divided into multiple subspaces based on the local directional features in our algorithm to achieve the optimal sparse representation for the image patches of different subsp...

In this paper, a novel method is proposed to identify the parameters of fractional-order systems. The proposed method converts the fractional differential equation to an algebraic one through a generalized operational matrix of block pulse functions. And thus, the output of the fractional system to be identified is represented by a matrix equation....

The sparse model often utilizes training samples to learn an over-complete dictionary, in order to obtain the redundant and sparse representation of signals. Designing simple, effective and flexible dictionary learning algorithms is one of the main and hot research topics in the information field. The dictionary learning methods based on synthesis...

Considering the disadvantage of the high complexity and ignoring signal's structural sparsity in Az.ast; Orthogonal Matching Pursuit (Az.ast;OMP) algorithm, a block Az.ast;OMP algorithm is proposed for block-sparse signals, and it is improved to solve the joint reconstruction problem for multiple signals in distributed compressed sensing. In the pr...

For most of those existing block-based compressed sensing of video, the same measurement matrix is usually utilized for all blocks, which underestimates the fact that the structural complexity and the movement varies from different regions. To address this issue, a novel block-based adaptive compressde sensing algorithm with variable sampling rate...

The conventional super-resolution algorithms on sparse representation reconstruct the high resolution image using one-stage high/low resolution dictionary pairs with inadequate detail information. In order to recover detail information as much as possible, two-stage dictionaries are explored in this paper. Then we train jointly multiple-frequency-b...

How to improve the reconstructed image quality using inherent prior knowledge of natural image is still a crucial issue in compressive imaging. In this paper, an efficient compressive imaging algorithm is proposed, which combines the sparse property of the entire image patches and the manifold property. In the algorithm, image patches are represent...

In wireless sensor networks (WSNs), energy-efficient data gathering and low-cost data transmission is very important for application, due to significant power constraints on the sensors. Our goal is to exploit temporal-spatial correlation and minimize the number of the required samples, reducing the cost of energy. We propose a data aggregation tec...

Compressed sensing uses the sparse prior of image representation, it can reconstruct image from samples much less than Nyquist rate. The sparse image representation and sparseness measure are two key ingredients which have important role on image reconstruction performance. To achieve the better sparse image representation, we split one octave scal...

Based on global sparse representation of image and local property of the patch, an efficient compressive imaging algorithm is proposed, which combined two priors: the low dimensional manifold property of local image patch and the sparse representation of analytic contourlet. The iterative hard threshold and manifold projection method are used to re...

Based on the standard compressed sensing, the model-based Compressed Sensing (CS) uses the tree structure priors, and solves the optimal reconstruction problem with two existing tree structure approximation which are greedy tree approximation and optimal tree approximation. Through numerous statistics test of wavelet relationship, a new tree struct...

The standard Compressed Sensing (CS) reconstructions of image exploit simply the sparse priors of the wavelet coefficients, ignoring the structural information of the wavelet coefficients. In this paper, the Hidden Markov Tree (HMT) model is integrated in the compressed sensing, which has been found successful in capturing the key features of the j...

The translation invariant analytic contourlet transform with low redundancy is proposed. In this transform, the circular symmetric filter banks decomposes image into multi-resolution detail subbands and one low-frequency subband, then the detail subbands are processed by Hilbert transform to generate two dimensional analytic signals. At last, the a...

For a natural image which includes both edge and texture information, the single basis function cannot reconstruct the image for compressed sensing optimally. In this paper, according to the Meyer's cartoon-texture model and biological vision function, the smooth and edge components are represented by Laplacian pyramid and circular symmetric contou...

The contourlet transform with anisotropy and directionality is a new extension to the wavelet transform. Because of its filter bank structure, the contourlet transform is not translation-invariant. In this paper, we propose the translation-invariant contourlet-like transform (TICLT) with lower redundancy than both the nonsubsampled contourlet trans...

In a wide variety of imaging applications (especially medical imaging), the theory of compressed sensing has shown it is surprisingly possible to reconstruct the entire original image from a partial set or subset of the Fourier transform of an image, if the image has a sparse or nearly sparse representation in some transform domain. Recently many f...

An algorithm for extracting fingerprint minutiae based on Harris corner detector is proposed. At the beginning, Harris corner detector is used to detect minutiae and high curvature dots in enhanced fingerprint image, subsequently, the postprocessing operations is utilized. In the post-processing operations, spurious minutiae are deleted base on the...

The circular symmetric contourlet transform (CSCT) overcomes the aliasing phenomenon of contourlet transform, and it has better direction selectivity than the contourlet transform. However, since the down-sampling operation in CSCT, it lacks the translation-invariant property which is essential for texture analysis. In this paper, the translation-i...

A rotation invariant texture classification algorithm based on dual-tree complex wavelet transform (DT-CWT) and support vector machines (SVM) is proposed. First, the texture image is transformed by Radon transform to convert the rotation to translation, the rotation invariant feature vector is composed of the energies of the subbands acquired by D...

The redundant contourlet transform implemented by undecimated pyramidal decomposition and directional filter bank is proposed. The circular symmetric filter bank satisfying perfect reconstruction conditions in the undecimated pyramidal decomposition is designed by McClellan transform. The adaptive local statistical model in the redundant contourlet...

The contourlet transform is a novel multiscale geometric analysis method. It can represent geometric features such as edges and texture more effectively than wavelet transform. In this paper, the circular symmetric contourlet transform (CSCT) which has similar frequency partition with cortex transform is proposed. In the CSCT, the circular symmetri...

Wavelet transform suffers with the limitation of poor directional selectivity. To overcome this disadvantage, a multidirectional and multiscale transform was proposed. First the circular symmetric filter bank decomposed the image into high frequency sub-band and low frequency subband, then the high frequency subband was further decomposed into dire...

The multi-directional and low-redundant wavelet with receptive fields properties of simple cells was constructed in 2D discrete Hilbert space. The multi-directional wavelet is named fanlet since its frequency spectrum support is fan-shaped. The fanlet transform can be implemented by circular symmetric multi-resolution decomposition and directional...