Guidance Image Research Articles

Fourier Domain Anomaly Detection and Spectral Fusion for Stripe Noise Removal of TIR Imagery

Stripe noise is a common and unwelcome noise pattern in various thermal infrared (TIR) image data including conventional TIR images and remote sensing TIR spectral images. Most existing stripe noise removal (destriping) methods are often difficult to keep a good and robust efficacy in dealing with the real-life complex noise cases. In this paper, based on the intrinsic spectral properties of TIR images and stripe noise, we propose a novel two-stage transform domain destriping method called Fourier domain anomaly detection and spectral fusion (ADSF). Considering the principal frequencies polluted by stripe noise as outliers in the statistical spectrum of TIR images, our naive idea is first to detect the potential anomalies and then correct them effectively in the Fourier domain to reconstruct a desired destriping result. More specifically, anomaly detection for stripe frequencies is achieved through a regional comparison between the original spectrum and the expected spectrum that statistically follows a generalized Laplacian regression model, and then an anomaly weight map is generated accordingly. In the correction stage, we propose a guidance-image-based spectrum fusion strategy, which integrates the original spectrum and the spectrum of a guidance image via the anomaly weight map. The final reconstruction result not only has no stripe noise but also maintains image structures and details well. Extensive real experiments are performed on conventional TIR images and remote sensing spectral images, respectively. The qualitative and quantitative assessment results demonstrate the superior effectiveness and strong robustness of the proposed method.

Noise suppression-guided image filtering for low-SNR CT reconstruction.

In practical computed tomography (CT) applications, projections with low signal-to-noise ratio (SNR) are often encountered due to the reduction of radiation dose or device limitations. In these situations, classical reconstruction algorithms, like simultaneous algebraic reconstruction technique (SART), cannot reconstruct high-quality CT images. Block-matching and 3D filtering (BM3D)-based iterative reconstruction algorithm (POCS-BM3D) has remarkable effect in dealing with CT reconstruction from noisy projections. However, BM3D may restrain noise with excessive loss of details in the case of low-SNR CT reconstruction. In order to achieve a preferable trade-off between noise suppression and edge preservation, we introduce guided image filtering (GIF) into low-SNR CT reconstruction, and propose noise suppression-guided image filtering reconstruction (NSGIFR) algorithm. In each iteration of NSGIFR, the output image of SART reserves more details and is used as input image of GIF, while the image denoised by BM3D serves as guidance image of GIF. Experimental results indicate that the proposed algorithm displays outstanding performance on preserving structures and suppressing noise for low-SNR CT reconstruction. NSGIFR can achieve more superior image quality than SART, POCS-TV and POCS-BM3D in terms of visual effect and quantitative analysis. Graphical abstract Block-matching and 3D filtering (BM3D)-based iterative reconstruction algorithm (POCS-BM3D) has remarkable effect in dealing with CT reconstruction from noisy projections. However, BM3D may restrain noise with excessive loss of details in the case of low-SNR CT reconstruction. In order to achieve a preferable trade-off between noise suppression and edge preservation, we introduce guided image filtering (GIF) into low-SNR CT reconstruction, and propose noise suppression-guided image filtering reconstruction (NSGIFR) algorithm.

Self‐guided filter for image denoising

The guided filter has been acknowledged as an exceptional edge-preserving filter whose output is a locally linear transform of the guidance image. However, the traditional guided filter heavily relies on the guidance image and fails to achieve the desired result when performing image denoising without a clear guidance image. In this study, to address this limitation, the authors propose a simple yet effective guided filter variant for the single image noise removing. They further show that the proposed denoising strategy can be easily realised by using the iterative framework. Moreover, the weak textured patches based image noise estimation is utilised to generate a clear intermediate image which makes the proposed method highly adaptable to the local noise level. Experimental results demonstrate that their proposed algorithm can compete with the state-of-the-art local denoising methods in edge-preserving.

A Non-Local Low-Rank Algorithm for Sub-Bottom Profile Sonar Image Denoising

Due to the influence of equipment instability and surveying environment, scattering echoes and other factors, it is sometimes difficult to obtain high-quality sub-bottom profile (SBP) images by traditional denoising methods. In this paper, a novel SBP image denoising method is developed for obtaining underlying clean images based on a non-local low-rank framework. Firstly, to take advantage of the inherent layering structures of the SBP image, a direction image is obtained and used as a guidance image. Secondly, the robust guidance weight for accurately selecting the similar patches is given. A novel denoising method combining the weight and a non-local low-rank filtering framework is proposed. Thirdly, after discussing the filtering parameter settings, the proposed method is tested in actual measurements of sub-bottom, both in deep water and shallow water. Experimental results validate the excellent performance of the proposed method. Finally, the proposed method is verified and compared with other methods quantificationally based on the synthetic images and has achieved the total average peak signal-to-noise ratio (PSNR) of 21.77 and structural similarity index (SSIM) of 0.573, which is far better than other methods.

Multivariate morphological reconstruction based fuzzy clustering with a weighting multi-channel guided image filter for color image segmentation

The fuzzy c-means clustering with guided image filter (GF) is a useful method for image segmentation. The single-channel GF can be efficiently applied to the gray-scale guidance image, but for the color guidance image, due to the high run-time overhead on the calculation of the inverse of the covariance matrix, it is a hard work to perform the multi-channel GF. To address this issue, we propose a novel weighting multi-channel guided image filter (WMGF) method. In this method, each channel of the color guidance image is utilized to guide the filtering for the input image independently and a novel weight is defined for each channel according to the variance of the image pixels in a local window, which greatly eliminates the mutual influence between different channels and brings about a low run-time overhead. In addition, based on the WMGF method, we present a new fuzzy c-means clustering algorithm ( $$\hbox {FCM}_{\scriptscriptstyle {WMGF }}$$ ) for the color image segmentation, in which the WMGF is performed on the membership matrix in each iteration of the fuzzy c-means clustering. To further enhance the different noise-immunity and edge preservation, the multivariate morphological reconstruction (MMR) method is introduced into the proposed fuzzy clustering method (MMR $$\_\hbox {FCM}_{\scriptscriptstyle {WMGF }}$$ ) to obtain higher segmentation precision. Experiments on color images with Salt & Pepper and Gaussian noises demonstrate the superiority of the proposed methods.

Fingerprint liveness detection based on guided filtering and hybrid image analysis

Fingerprints are widely used for biometric recognition. However, many spoofing attacks based on an artificially made fingerprint occur. In this study, the authors propose an approach to detect fingerprint liveness which uses the guided filtering and hybrid image analysis. This study deals with the problem of ignoring the contribution that is brought by the sharp features when analysing the denoised image. The method described utilises both the enhanced sharp features and denoised features from the hybrid images to get better results. The input fingerprint is pre-processed by region of interest extraction and then is filtered by a guidance image for obtaining the denoised image. Then, histogram equalisation is introduced to eliminate the impact of illumination condition. The authors extract the co-occurrence of adjacent local binary pattern features from both the cropped images and the denoised images. Whilst concatenating both the features together to form a long feature, t-Distributed Stochastic Neighbour Embedding is applied to reduce the data dimension. The authors consider the fingerprint liveness detection as a two-class classification problem and use support vector machine with radial basis function kernel to solve this problem. The authors evaluate the experiments on three benchmark data sets. Experimental results demonstrate that the accuracy of the proposed method can outperform most of the state-of-art methods.

Sparse Features with Fast Guided Filtering for Medical Image Fusion

In medical image fusion, partially focused images are combined into a completely informative in focus image of the same scene, helping the clinical applicability of medical image for assessment on sports injuries diagnosis and rehabilitation. However, many state-of-the-art fusion methods seldom consider structural differences between guidance and input images, so not all the significant features of the source images can be well preserved for a completely informative focused image. In this paper, a fusion method via fast guided filtering with sparse features is proposed. The source images are decomposed into base and detail layers using an average filter. They are also split into low-rank and sparse parts by solving convex programs. Saliency maps are constructed based on the sparse features using low pass and high pass filters. Fast guided filtering is used to optimize saliency maps for constructing weighted maps of the base and detail layers, and for maintaining the spatial consistency between the source images and the layers. The fused base and detail layers are integrated to construct the final fused image. Experiments on nine pair of medical images demonstrate that the proposed method achieved better medical images compared to other reported methods in terms of both subjective and objective evaluations.

Ultrasound-guided Needle nerve contact positions detection in Regional Anaesthesia using Needle tip force and Opening injection pressure

Ultrasound imaging guidance has been widely used in peripheral nerve block to guide the anaesthetist to insert the needle and observe injection fluid diffusion. However, needle-nerve contact(NNC) positions visualization is difficult to achieve because of requirement of high image solution and expert. To better guide the anaesthetist to predict the distance from needle to nerve and reduce the needle trauma, we attempted to use Needle tip force(NTF) and Opening injection pressure (OIP) to identify the NNC positions under ultrasound guidance. The relationship between OIP and NNC, NTF and NNC, NTF and OIP is unclear now, and we conducted a prospective, observational study to define the relationship. In this study, we designed a multi-function block needle and built a real-time NTF and OIP measurement which could trace and record NTF and OIP data in the simulator sciatic nerve. NNC positions were divided into three distinct NNC positions: Pre-NNC, First-NNC, Forceful-NNC using ultrasound guidance image. We statistically analysis NTF, OIP in the three NNC positions, and the results indicate the NTF differentiate in three NNC positions(p<0.001), OIP differentiate as well (p<0.001) and the NTF has a highly positive with the OIP in the same NNC positions with Pearson correlation 0.62.

A Multi-Scale Superpixel-Guided Filter Feature Extraction and Selection Approach for Classification of Very-High-Resolution Remotely Sensed Imagery

In this article, a novel feature selection-based multi-scale superpixel-based guided filter (FS-MSGF) method for classification of very-high-resolution (VHR) remotely sensed imagery is proposed. Improved from the original guided filter (GF) algorithm used in the classification, the guidance image in the proposed approach is constructed based on the superpixel-level segmentation. By taking into account the object boundaries and the inner-homogeneity, the superpixel-level guidance image leads to the geometrical information of land-cover objects in VHR images being better depicted. High-dimensional multi-scale guided filter (MSGF) features are then generated, where the multi-scale information of those land-cover classes is better modelled. In addition, for improving the computational efficiency without the loss of accuracy, a subset of those MSGF features is then automatically selected by using an unsupervised feature selection method, which contains the most distinctive information in all constructed MSGF features. Quantitative and qualitative classification results obtained on two QuickBird remotely sensed imagery datasets covering the Zurich urban scene are provided and analyzed, which demonstrate that the proposed methods outperform the state-of-the-art reference techniques in terms of higher classification accuracies and higher computational efficiency.

Spatial-Spectral Joint Classification of Hyperspectral Image With Locality and Edge Preserving

Hyperspectral image (HSI) classification is an important part of its processing and application. Aiming at the problems of high data dimensionality and high spatial neighborhood correlation in HSI classification, we propose a spatial-spectral joint classification method of HSI with locality and edge preserving in this article. First, the input HSI is normalized, and the feature is extracted by principal component analysis. The first principal component image is taken as the guidance image. Second, guided filtering is used to extract the spatial features of each band separately. Then, the extracted spatial features are superimposed, and low-dimensional embedding is completed through local Fisher discriminant analysis. Finally, the obtained low-dimensional embedded features are input into a random forest classifier to get classification results. The experimental results of two HSI show that the proposed method achieves higher classification accuracy than other related methods. In the case of randomly selecting 10% and 1% samples from each class of ground object as training samples, the overall classification accuracy is improved to 99.57% and 97.79%, respectively. This method effectively uses the spatial and local information of the image in low dimensional embedding, and preserves the boundaries of the ground objects, thus improving the classification effect.

基于引导图像和自适应支持域的立体匹配

基于引导图像和自适应支持域的立体匹配

An Improved Feature Set for Hyperspectral Image Classification: Harmonic Analysis Optimized by Multiscale Guided Filter

Effective features derived from an original hyperspectral image (HSI) are quite important to improve the classification performance. An improved feature set, namely HGFM, is constructed by integrating harmonic analysis (HA) optimized by a multiscale guided filter (GF) with morphological operation for HSI classification. To establish HGFM, HA is first adopted to convert the HSI from spectral space to the frequency domain represented by amplitude, phase, and residual. With the first component of minimum noise fraction obtained from the original HSI as the guidance image, the harmonic components are then processed by the multiscale GF. Finally, the obtained results are then operated via morphological opening by reconstruction and closing by reconstruction to generate an improved feature set for classification. The HGFM features are input to an ensemble learning (EL) based on classification framework, in which EL plays an auxiliary role to enhance the classification stability and reliability. Three commonly used HSIs are used for experiments, and different feature sets are evaluated by comparing EL and rotation forest, support vector machine optimized by particle swarm optimization, random forest, and others. Compared with benchmark feature sets, the proposed HGFM feature set can better depict the details of objects easily, and the experimental results confirm the effectiveness in terms of classification accuracy and generalization ability.

Filter-Based Landweber Iterative Method for Reconstructing the Light Field

The light field reconstruction aims to recover the 4D light field from the 3D focal stack, so it is a severe incomplete data reconstructing problem. The Landweber iteration is a general method to find the least-square solution of the discretized linear system, which is widely used for solving such problem. The relaxation strategies of the Landweber iteration are essential for both accelerating iterative convergence and decreasing errors. As a fast and non-approximate linear-time method, the fast guided filtering has good edge-preserving smoothing properties. It can be incorporated into the light field reconstruction process to improve the precision of the reconstructed results. In this paper, we propose a filter-based Landweber iterative method by introducing the optimized relaxation strategy and the fast guided filter. Specifically, the updated image in the iteration step contains the guidance image and the image obtained by the Landweber iteration, which makes the proposed method flexible. The experimental results show that the proposed method is more practical and effective compared with the relevant reconstruction methods.

Deep Guidance Network for Biomedical Image Segmentation

Segmentation of 2D images is a fundamental problem for biomedical image analysis. The most widely used architecture for biomedical image segmentation is U-Net. U-Net introduces skip-connections to restore the spatial information loss caused by down-sampling operations. However, for some tasks such as the retinal vessel segmentation, the loss information of structure can not be fully recovered since the vessels is merely a curve line that can not be detected after several convolutions. In this paper, we introduce a deep guidance network to segment the biomedical image. Our proposed network consists of a guided image filter module to restore the structure information through the guidance image. Our method enables end to end training and fast inference (43ms for one image). We conduct extensive experiments for the task of vessel segmentation and optic disc and cup segmentation. The experiments on four publicly available datasets: ORIGA, REFUGE, DRIVE, and CHASEDB1 verify the effectiveness of our method.

A Novel UAV Sensing Image Defogging Method

Imaging of unmanned aerial vehicle easily suffer from haze, resulting in decline in the quality of required remote sensing images. The influence brings great challenges in later analysis and process. Although dark channel prior has acquired substantial achievements, some limitations, including imprecise estimation of atmospheric light, color distortion, and lower brightness of defogging image, still exist. In this article, to target these drawbacks, a novel defogging method for single image is proposed. First, a novel atmospheric scattering model is proposed to define the more accurate atmospheric light by introducing an adaptive variable strategy. Next, unlike traditional dark channel prior, a novel estimation method is presented by fusing dark and light channels to estimate more precise atmospheric light and transmittance. Then, we adopt the gray image corresponding to color image as a guidance image to refine the transmittance to further decrease the time complexity. Aiming at the region of low transmittance, a novel compensation function is created to improve the region of low transmittance and avoid color distortion. Moreover, a simple and effective calculation method is proposed to determine parameters in compensation function. Finally, the clear remote sensing image is established by an improved atmospheric scattering model. Extensive experiments on real-world datasets demonstrate that the proposed method outperforms several other state-of-the-art approaches both on subjective and objective quality evaluations.

Guidance Based Improved Depth Upsampling With Better Initial Estimate

Like optical images, depth images are also gaining popularity because of its use in many applications like robot navigation, augmented reality, 3DTV and more. The commercially available depth cameras generate depth images which suffer from low spatial resolution, corrupted with noise, and missing regions. Such images need to be super-resolved, denoised and inpainted before using them to have better accuracy. Super-resolution (SR) techniques can be used to produce a high-resolution output. Since SR is an ill-posed inverse problem, a good initial estimate is always a good regulariser to find the optimal solution. We propose an initial estimate as part of our SR pipeline, esp. ×8, which will helps in quick convergence and accurate output. We propose a cascade approach by combining residual interpolation (RI) method with anisotropic total generalised variation (ATGV) method, both uses HR guidance image. The improvements are shown qualitative and quantitative with different levels of noise.

GPU acceleration of liver enhancement for tumor segmentation

Background and objective: Medical image segmentation plays a vital role in medical image analysis. There are many algorithms developed for medical image segmentation which are based on edge or region characteristics. These are dependent on the quality of the image. The contrast of a CT or MRI image plays an important role in identifying region of interest i.e. lesion(s). In order to enhance the contrast of image, clinicians generally use manual histogram adjustment technique which is based on 1D histogram specification. This is time consuming and results in poor distribution of pixels over the image. Cross modality based contrast enhancement is 2D histogram specification technique. This is robust and provides a more uniform distribution of pixels over CT image by exploiting the inner structure information from MRI image. This helps in increasing the sensitivity and accuracy of lesion segmentation from enhanced CT image. The sequential implementation of cross modality based contrast enhancement is slow. Hence we propose GPU acceleration of cross modality based contrast enhancement for tumor segmentation.Methods: The aim of this study is fast parallel cross modality based contrast enhancement for CT liver images. This includes pairwise 2D histogram, histogram equalization and histogram matching. The sequential implementation of the cross modality based contrast enhancement is computationally expensive and hence time consuming. We propose persistence and grid-stride loop based fast parallel contrast enhancement for CT liver images. We use enhanced CT liver image for the lesion or tumor segmentation. We implement the fast parallel gradient based dynamic seeded region growing for lesion segmentation.Results: The proposed parallel approach is 104.416 ( ± 5.166) times faster compared to the sequential implementation and increases the sensitivity and specificity of tumor segmentation.Conclusion: The cross modality approach is inspired by 2D histogram specification which incorporates spatial information existing in both guidance and input images for remapping the input image intensity values. The cross modality based liver contrast enhancement improves the quality of tumor segmentation.

Target tracking from infrared imagery via an improved appearance model

Moving target tracking from the infrared images is a challenging task due to the targets’ unstable appearance, complex background clutters and a limited number of pixels for each target. To address this challenge, we propose an infrared target tracking method that combines the Sparsity-based Generative Model (SGM) and Guided Filter-based Discriminative Classifier (GFDC) with the particle filter. The SGM model extracts an effective template histogram, which considers the spatial information of each image patch. In the GFDC model, the guided filter uses the structural information of the guidance image in order to smooth the input image, while preserving the edges. This enhances the information of templates while suppressing the noise in the image data. The Bayesian classification is then applied to take the background into consideration when the model is fused. The online update scheme not only considers appearance variations but also relieves the tracking deviation problem. At last, we utilize the joint confidence values to measure the candidate templates for the target position estimation. Experiments based on infrared video sequences have validated the superior tracking performance of the proposed method.

Hyperspectral image recovery based on fusion of coded aperture snapshot spectral imaging and RGB images by guided filtering

In compressive hyperspectral imaging systems, the key issue is to accurately and efficiently recover 3D spectral data cubes from spectrally undersampled 2D projections. In this paper, a guided-filtering-based reconstruction algorithm is proposed to recover hyperspectral data cube with a joint coded aperture snapshot spectral imaging (CASSI) and RGB imaging system. In order to preserve the spectral and spatial details in the CASSI and the RGB branch to the greatest extent, 2D projections from the two branches are reconstructed into hyperspectral images individually with a pre-trained dictionary. The appropriate combination of the information from the two branches is realized through guided image filtering, which is applied on the reconstruction result of the CASSI branch with the result from the RGB branch as the guidance image. In virtue of combination between the well-performed spectral fidelity of the CASSI branch and the high spatial accuracy of the RGB branch, the proposed guided-filtering-based compressive hyperspectral imaging (GFCHI) not only overcomes the block effect and blurring common in the state-of-the-art methods, but also improves spectral fidelity related to key applications of hyperspectral imaging. The quantitative metrics associated with spatial, spectral and overall performance demonstrate the state-of-the-art reconstruction quality of GFCHI.

Infrared and Visible Image Fusion through Details Preservation.

In many actual applications, fused image is essential to contain high-quality details for achieving a comprehensive representation of the real scene. However, existing image fusion methods suffer from loss of details because of the error accumulations of sequential tasks. This paper proposes a novel fusion method to preserve details of infrared and visible images by combining new decomposition, feature extraction, and fusion scheme. For decomposition, different from the most decomposition methods by guided filter, the guidance image contains only the strong edge of the source image but no other interference information so that rich tiny details can be decomposed into the detailed part. Then, according to the different characteristics of infrared and visible detail parts, a rough convolutional neural network (CNN) and a sophisticated CNN are designed so that various features can be fully extracted. To integrate the extracted features, we also present a multi-layer features fusion strategy through discrete cosine transform (DCT), which not only highlights significant features but also enhances details. Moreover, the base parts are fused by weighting method. Finally, the fused image is obtained by adding the fused detail and base part. Different from the general image fusion methods, our method not only retains the target region of source image but also enhances background in the fused image. In addition, compared with state-of-the-art fusion methods, our proposed fusion method has many advantages, including (i) better visual quality of fused-image subjective evaluation, and (ii) better objective assessment for those images.