Unwrapping phase contrast MRI by iterative graph cuts.

PURPOSE: To develop and evaluate a phase unwrapping method for cine phase contrast MRI based on graph cuts. METHODS: A proposed Iterative Graph Cuts method was evaluated in 10 cardiac patients with two-dimensional flow quantification which was repeated at low venc settings to provoke wrapping. The images were also unwrapped by a path-following method (ROMEO), and a Laplacian-based method (LP). Net flow was quantified using semi-automatic vessel segmentation. High venc images were also wrapped retrospectively to asses the residual amount of wrapped voxels. RESULTS: The absolute net flow error after unwrapping at venc = 100 cm/s was 1.8 mL, which was 0.83 mL smaller than for LP. The repeatability error at high venc without unwrapping was 2.5 mL. The error at venc = 50 cm/s was 7.5 mL, which was 8.2 mL smaller than for ROMEO and 5.7 mL smaller than for LP. For retrospectively wrapped images with synthetic venc of 100/50/25 cm/s, the residual amount of wrapped voxels was 0.00/0.12/0.79%, which was 0.09/0.26/8.0 percentage points smaller than for LP. With synthetic venc of 25 cm/s, omitting magnitude information resulted in 3.2 percentage points more wrapped voxels, and only spatial/temporal unwrapping resulted in 4.6/21 percentage points more wrapped voxels compared to spatiotemporal unwrapping. CONCLUSION: Iterative Graph Cuts enables unwrapping of cine phase contrast MRI with very small errors, except for at extreme blood velocities, with equal or better performance compared to ROMEO and LP. The use of magnitude information and spatiotemporal unwrapping is recommended.

Improved body quantitative susceptibility mapping by using a variable-layer single-min-cut graph-cut for field-mapping.

PURPOSE: To develop a robust algorithm for field-mapping in the presence of water-fat components, large B0 field inhomogeneities and MR signal voids and to apply the developed method in body applications of quantitative susceptibility mapping (QSM). METHODS: A framework solving the cost-function of the water-fat separation problem in a single-min-cut graph-cut based on the variable-layer graph construction concept was developed. The developed framework was applied to a numerical phantom enclosing an MR signal void, an air bubble experimental phantom, 14 large field of view (FOV) head/neck region in vivo scans and to 6 lumbar spine in vivo scans. Field-mapping and subsequent QSM results using the proposed algorithm were compared to results using an iterative graph-cut algorithm and a formerly proposed single-min-cut graph-cut. RESULTS: The proposed method was shown to yield accurate field-map and susceptibility values in all simulation and in vivo datasets when compared to reference values (simulation) or literature values (in vivo). The proposed method showed improved field-map and susceptibility results compared to iterative graph-cut field-mapping especially in regions with low SNR, strong field-map variations and high R2∗ values. CONCLUSIONS: A single-min-cut graph-cut field-mapping method with a variable-layer construction was developed for field-mapping in body water-fat regions, improving quantitative susceptibility mapping particularly in areas close to MR signal voids.

GoRG: Towards a GPU-Accelerated Multiview Hyperspectral Depth Estimation Tool for Medical Applications.

HyperSpectral (HS) images have been successfully used for brain tumor boundary detection during resection operations. Nowadays, these classification maps coexist with other technologies such as MRI or IOUS that improve a neurosurgeon's action, with their incorporation being a neurosurgeon's task. The project in which this work is framed generates an unified and more accurate 3D immersive model using HS, MRI, and IOUS information. To do so, the HS images need to include 3D information and it needs to be generated in real-time operating room conditions, around a few seconds. This work presents Graph cuts Reference depth estimation in GPU (GoRG), a GPU-accelerated multiview depth estimation tool for HS images also able to process YUV images in less than 5.5 s on average. Compared to a high-quality SoA algorithm, MPEG DERS, GoRG YUV obtain quality losses of -0.93 dB, -0.6 dB, and -1.96% for WS-PSNR, IV-PSNR, and VMAF, respectively, using a video synthesis processing chain. For HS test images, GoRG obtains an average RMSE of 7.5 cm, with most of its errors in the background, needing around 850 ms to process one frame and view. These results demonstrate the feasibility of using GoRG during a tumor resection operation.

Iterative Min Cut Clustering Based on Graph Cuts.

Clustering nonlinearly separable datasets is always an important problem in unsupervised machine learning. Graph cut models provide good clustering results for nonlinearly separable datasets, but solving graph cut models is an NP hard problem. A novel graph-based clustering algorithm is proposed for nonlinearly separable datasets. The proposed method solves the min cut model by iteratively computing only one simple formula. Experimental results on synthetic and benchmark datasets indicate the potential of the proposed method, which is able to cluster nonlinearly separable datasets with less running time.

A dictionary-based graph-cut algorithm for MRI reconstruction.

PURPOSE: Compressive sensing (CS)-based image reconstruction methods have proposed random undersampling schemes that produce incoherent, noise-like aliasing artifacts, which are easier to remove. The denoising process is critically assisted by imposing sparsity-enforcing priors. Sparsity is known to be induced if the prior is in the form of the Lp (0 ≤ p ≤ 1) norm. CS methods generally use a convex relaxation of these priors such as the L1 norm, which may not exploit the full power of CS. An efficient, discrete optimization formulation is proposed, which works not only on arbitrary Lp -norm priors as some non-convex CS methods do, but also on highly non-convex truncated penalty functions, resulting in a specific type of edge-preserving prior. These advanced features make the minimization problem highly non-convex, and thus call for more sophisticated minimization routines. THEORY AND METHODS: The work combines edge-preserving priors with random undersampling, and solves the resulting optimization using a set of discrete optimization methods called graph cuts. The resulting optimization problem is solved by applying graph cuts iteratively within a dictionary, defined here as an appropriately constructed set of vectors relevant to brain MRI data used here. RESULTS: Experimental results with in vivo data are presented. CONCLUSION: The proposed algorithm produces better results than regularized SENSE or standard CS for reconstruction of in vivo data.

ALS Point Cloud Classification by Integrating an Improved Fully Convolutional Network into Transfer Learning with Multi-Scale and Multi-View Deep Features.

Airborne laser scanning (ALS) point cloud has been widely used in various fields, for it can acquire three-dimensional data with a high accuracy on a large scale. However, due to the fact that ALS data are discretely, irregularly distributed and contain noise, it is still a challenge to accurately identify various typical surface objects from 3D point cloud. In recent years, many researchers proved better results in classifying 3D point cloud by using different deep learning methods. However, most of these methods require a large number of training samples and cannot be widely used in complex scenarios. In this paper, we propose an ALS point cloud classification method to integrate an improved fully convolutional network into transfer learning with multi-scale and multi-view deep features. First, the shallow features of the airborne laser scanning point cloud such as height, intensity and change of curvature are extracted to generate feature maps by multi-scale voxel and multi-view projection. Second, these feature maps are fed into the pre-trained DenseNet201 model to derive deep features, which are used as input for a fully convolutional neural network with convolutional and pooling layers. By using this network, the local and global features are integrated to classify the ALS point cloud. Finally, a graph-cuts algorithm considering context information is used to refine the classification results. We tested our method on the semantic 3D labeling dataset of the International Society for Photogrammetry and Remote Sensing (ISPRS). Experimental results show that overall accuracy and the average F1 score obtained by the proposed method is 89.84% and 83.62%, respectively, when only 16,000 points of the original data are used for training.

Two-Layered Graph-Cuts-Based Classification of LiDAR Data in Urban Areas.

Classifying the LiDAR (Light Detection and Ranging) point cloud in the urban environment is a challenging task. Due to the complicated structures of urban objects, it is difficult to find suitable features and classifiers to efficiently category the points. A two-layered graph-cuts-based classification framework is addressed in this study. The hierarchical framework includes a bottom layer that defines the features and classifies point clouds at the point level as well as a top layer that defines the features and classifies the point cloud at the object level. A novel adaptive local modification method is employed to model the interactions between these two layers. The iterative graph cuts algorithm runs around the bottom and top layers to optimize the classification. In this way, the addressed framework benefits from the integration of point features and object features to improve the classification. The experiments demonstrate that the proposed method is capable of producing classification results with high accuracy and efficiency.

Dual echo Dixon imaging with a constrained phase signal model and graph cuts reconstruction.

PURPOSE: The purpose of this work is to derive and demonstrate constrained-phase dual-echo Dixon imaging within a maximum likelihood framework solved with a regularized graph-cuts-guided optimization. THEORY AND METHODS: Dual-echo Dixon reconstruction is fundamentally underdetermined; however, adopting a constrained-phase signal model reduces the number of unknowns and the nonlinear problem can be solved under a maximum likelihood framework. Period shifts in the field map (manifesting as fat/water signal swaps) must also be corrected. Here, a regularized cost function promotes a smooth field map and is solved with a graph-cuts-guided greedy binary optimization. The reconstruction shown here is compared to two other prevalent Dixon reconstructions in experimental phantom and human studies. RESULTS: Reconstructed images of the water and fat signal are shown for a phantom study, and in vivo studies of foot/ankle, pelvis, and CE-MRA of the thighs. The method shown here compared favorably with the other two methods. Large field inhomogeneities on the order of 20 ppm were resolved, thereby avoiding the fat and water signal swaps present in images reconstructed with the other methods. CONCLUSION: Constrained-phase dual-echo Dixon imaging solved with a regularized graph-cuts-guided optimization has been derived and demonstrated to successfully separate water and fat images in the presence of large magnetic field inhomogeneities. Magn Reson Med 78:2203-2215, 2017. © 2017 International Society for Magnetic Resonance in Medicine.

Phase unwrapping with graph cuts optimization and dual decomposition acceleration for 3D high-resolution MRI data.

PURPOSE: Existence of low SNR regions and rapid-phase variations pose challenges to spatial phase unwrapping algorithms. Global optimization-based phase unwrapping methods are widely used, but are significantly slower than greedy methods. In this paper, dual decomposition acceleration is introduced to speed up a three-dimensional graph cut-based phase unwrapping algorithm. METHODS: The phase unwrapping problem is formulated as a global discrete energy minimization problem, whereas the technique of dual decomposition is used to increase the computational efficiency by splitting the full problem into overlapping subproblems and enforcing the congruence of overlapping variables. Using three dimensional (3D) multiecho gradient echo images from an agarose phantom and five brain hemorrhage patients, we compared this proposed method with an unaccelerated graph cut-based method. RESULTS: Experimental results show up to 18-fold acceleration in computation time. CONCLUSIONS: Dual decomposition significantly improves the computational efficiency of 3D graph cut-based phase unwrapping algorithms. Magn Reson Med 77:1353-1358, 2017. © 2016 International Society for Magnetic Resonance in Medicine.

Multi-scale graph-cut algorithm for efficient water-fat separation.

PURPOSE: To improve the accuracy and robustness to noise in water-fat separation by unifying the multiscale and graph cut based approaches to B0 -correction. METHODS: A previously proposed water-fat separation algorithm that corrects for B0 field inhomogeneity in 3D by a single quadratic pseudo-Boolean optimization (QPBO) graph cut was incorporated into a multi-scale framework, where field map solutions are propagated from coarse to fine scales for voxels that are not resolved by the graph cut. The accuracy of the single-scale and multi-scale QPBO algorithms was evaluated against benchmark reference datasets. The robustness to noise was evaluated by adding noise to the input data prior to water-fat separation. RESULTS: Both algorithms achieved the highest accuracy when compared with seven previously published methods, while computation times were acceptable for implementation in clinical routine. The multi-scale algorithm was more robust to noise than the single-scale algorithm, while causing only a small increase (+10%) of the reconstruction time. CONCLUSION: The proposed 3D multi-scale QPBO algorithm offers accurate water-fat separation, robustness to noise, and fast reconstruction. The software implementation is freely available to the research community. Magn Reson Med 78:941-949, 2017. © 2016 International Society for Magnetic Resonance in Medicine.

Rapid Texture Optimization of Three-Dimensional Urban Model Based on Oblique Images.

Seamless texture mapping is one of the key technologies for photorealistic 3D texture reconstruction. In this paper, a method of rapid texture optimization of 3D urban reconstruction based on oblique images is proposed aiming at the existence of texture fragments, seams, and inconsistency of color in urban 3D texture mapping based on low-altitude oblique images. First, we explore implementing radiation correction on the experimental images with a radiation procession algorithm. Then, an efficient occlusion detection algorithm based on OpenGL is proposed according to the mapping relation between the terrain triangular mesh surface and the images to implement the occlusion detection of the visible texture on the triangular facets as well as create a list of visible images. Finally, a texture clustering algorithm is put forward based on Markov Random Field utilizing the inherent attributes of the images and solve the energy function minimization by Graph-Cuts. The experimental results display that the method is capable of decreasing the existence of texture fragments, seams, and inconsistency of color in the 3D texture model reconstruction.

Seamline Determination Based on PKGC Segmentation for Remote Sensing Image Mosaicking.

This paper presents a novel method of seamline determination for remote sensing image mosaicking. A two-level optimization strategy is applied to determine the seamline. Object-level optimization is executed firstly. Background regions (BRs) and obvious regions (ORs) are extracted based on the results of parametric kernel graph cuts (PKGC) segmentation. The global cost map which consists of color difference, a multi-scale morphological gradient (MSMG) constraint, and texture difference is weighted by BRs. Finally, the seamline is determined in the weighted cost from the start point to the end point. Dijkstra's shortest path algorithm is adopted for pixel-level optimization to determine the positions of seamline. Meanwhile, a new seamline optimization strategy is proposed for image mosaicking with multi-image overlapping regions. The experimental results show the better performance than the conventional method based on mean-shift segmentation. Seamlines based on the proposed method bypass the obvious objects and take less time in execution. This new method is efficient and superior for seamline determination in remote sensing image mosaicking.

Supervoxels for Graph Cuts-Based Deformable Image Registration Using Guided Image Filtering.

In this work we propose to combine a supervoxel-based image representation with the concept of graph cuts as an efficient optimization technique for 3D deformable image registration. Due to the pixels/voxels-wise graph construction, the use of graph cuts in this context has been mainly limited to 2D applications. However, our work overcomes some of the previous limitations by posing the problem on a graph created by adjacent supervoxels, where the number of nodes in the graph is reduced from the number of voxels to the number of supervoxels. We demonstrate how a supervoxel image representation, combined with graph cuts-based optimization can be applied to 3D data. We further show that the application of a relaxed graph representation of the image, followed by guided image filtering over the estimated deformation field, allows us to model 'sliding motion'. Applying this method to lung image registration, results in highly accurate image registration and anatomically plausible estimations of the deformations. Evaluation of our method on a publicly available Computed Tomography lung image dataset (www.dir-lab.com) leads to the observation that our new approach compares very favorably with state-of-the-art in continuous and discrete image registration methods achieving Target Registration Error of 1.16mm on average per landmark.

Automated cerebellar lobule segmentation with application to cerebellar structural analysis in cerebellar disease.

The cerebellum plays an important role in both motor control and cognitive function. Cerebellar function is topographically organized and diseases that affect specific parts of the cerebellum are associated with specific patterns of symptoms. Accordingly, delineation and quantification of cerebellar sub-regions from magnetic resonance images are important in the study of cerebellar atrophy and associated functional losses. This paper describes an automated cerebellar lobule segmentation method based on a graph cut segmentation framework. Results from multi-atlas labeling and tissue classification contribute to the region terms in the graph cut energy function and boundary classification contributes to the boundary term in the energy function. A cerebellar parcellation is achieved by minimizing the energy function using the α-expansion technique. The proposed method was evaluated using a leave-one-out cross-validation on 15 subjects including both healthy controls and patients with cerebellar diseases. Based on reported Dice coefficients, the proposed method outperforms two state-of-the-art methods. The proposed method was then applied to 77 subjects to study the region-specific cerebellar structural differences in three spinocerebellar ataxia (SCA) genetic subtypes. Quantitative analysis of the lobule volumes shows distinct patterns of volume changes associated with different SCA subtypes consistent with known patterns of atrophy in these genetic subtypes.

Automated analysis of hip joint cartilage combining MR T2 and three-dimensional fast-spin-echo images.

PURPOSE: To validate a fully automated scheme to extract biochemical information from the hip joint cartilages using MR T2 mapping images incorporating segmentation of co-registered three-dimensional Fast-Spin-Echo (3D-SPACE) images. METHODS: Manual analyses of unilateral hip (3 Tesla) MR images of 24 asymptomatic volunteers were used to validate a 3D deformable model method for automated cartilage segmentation of SPACE scans, partitioning of the individual femoral and acetabular cartilage plates into clinically defined sub-regions and propagating these results to T2 maps to calculate region-wise T2 value statistics. Analyses were completed on a desktop computer (∼ 10 min per case). RESULTS: The mean voxel overlap between automated A and manual M segmentations of the cartilage volumes in the (clinically based) SPACE images was 73% (100 × 2|A∩M|/[|A|+|M|]). The automated and manual analyses demonstrated a relative difference error <10% in the median "T2 average signal" for each cartilage plate. The automated and manual analyses showed consistent patterns between significant differences in T2 data across the hip cartilage sub-regions. CONCLUSION: The good agreement between the manual and automatic analyses of T2 values indicates the use of structural 3D-SPACE MR images with the proposed method provides a promising approach for automated quantitative T2 assessment of hip joint cartilages.

A Unified Framework for Street-View Panorama Stitching.

In this paper, we propose a unified framework to generate a pleasant and high-quality street-view panorama by stitching multiple panoramic images captured from the cameras mounted on the mobile platform. Our proposed framework is comprised of four major steps: image warping, color correction, optimal seam line detection and image blending. Since the input images are captured without a precisely common projection center from the scenes with the depth differences with respect to the cameras to different extents, such images cannot be precisely aligned in geometry. Therefore, an efficient image warping method based on the dense optical flow field is proposed to greatly suppress the influence of large geometric misalignment at first. Then, to lessen the influence of photometric inconsistencies caused by the illumination variations and different exposure settings, we propose an efficient color correction algorithm via matching extreme points of histograms to greatly decrease color differences between warped images. After that, the optimal seam lines between adjacent input images are detected via the graph cut energy minimization framework. At last, the Laplacian pyramid blending algorithm is applied to further eliminate the stitching artifacts along the optimal seam lines. Experimental results on a large set of challenging street-view panoramic images captured form the real world illustrate that the proposed system is capable of creating high-quality panoramas.

Semi-automatic breast ultrasound image segmentation based on mean shift and graph cuts.

Computerized tumor segmentation on breast ultrasound (BUS) images remains a challenging task. In this paper, we proposed a new method for semi-automatic tumor segmentation on BUS images using Gaussian filtering, histogram equalization, mean shift, and graph cuts. The only interaction required was to select two diagonal points to determine a region of interest (ROI) on an input image. The ROI image was shrunken by a factor of 2 using bicubic interpolation to reduce computation time. The shrunken image was smoothed by a Gaussian filter and then contrast-enhanced by histogram equalization. Next, the enhanced image was filtered by pyramid mean shift to improve homogeneity. The object and background seeds for graph cuts were automatically generated on the filtered image. Using these seeds, the filtered image was then segmented by graph cuts into a binary image containing the object and background. Finally, the binary image was expanded by a factor of 2 using bicubic interpolation, and the expanded image was processed by morphological opening and closing to refine the tumor contour. The method was implemented with OpenCV 2.4.3 and Visual Studio 2010 and tested for 38 BUS images with benign tumors and 31 BUS images with malignant tumors from different ultrasound scanners. Experimental results showed that our method had a true positive rate (TP) of 91.7%, a false positive (FP) rate of 11.9%, and a similarity (SI) rate of 85.6%. The mean run time on Intel Core 2.66 GHz CPU and 4 GB RAM was 0.49 ± 0.36 s. The experimental results indicate that the proposed method may be useful in BUS image segmentation.

Complex carotid artery segmentation in multi-contrast MR sequences by improved optimal surface graph cuts based on flow line learning.

Carotid atherosclerosis is one of the leading causes of cardiovascular disease with high mortality. Multi-contrast MRI can identify atherosclerotic plaque components with high sensitivity and specificity. Accurate segmentation of the diseased carotid artery from MR images is very essential to quantitatively evaluate the state of atherosclerosis. However, due to the complex morphology of atherosclerosis plaques and the lack of well-annotated data, the segmentation of lumen and wall is very challenging. Different from popular deep learning methods, in this paper, we propose an integration segmentation framework by introducing a lightweight prediction model and improved optimal surface graph cuts (OSG), which adopts a simplified flow line sampling and post-reconstructing method to reduce the cost of graph construction. Moreover, a flexibly adaptive smoothing penalty is presented for maintaining the shape of diseased carotid surface. For the experiments, we have collected an MR image dataset from patients with carotid atherosclerosis and evaluated our method by cross-validation. It can reach 89.68%/80.29% of dice coefficients and 0.2480 mm/0.3396 mm of average surface distances on the lumen/wall segmentation, respectively. The experimental results show that our method can generate precise and reliable segmentation of both lumen and wall of diseased carotid artery with a quite small training cost.

Combining global climate models using graph cuts.

Global Climate Models are the main tools for climate projections. Since many models exist, it is common to use Multi-Model Ensembles to reduce biases and assess uncertainties in climate projections. Several approaches have been proposed to combine individual models and extract a robust signal from an ensemble. Among them, the Multi-Model Mean (MMM) is the most commonly used. Based on the assumption that the models are centered around the truth, it consists in averaging the ensemble, with the possibility of using equal weights for all models or to adjust weights to favor some models. In this paper, we propose a new alternative to reconstruct multi-decadal means of climate variables from a Multi-Model Ensemble, where the local performance of the models is taken into account. This is in contrast with MMM where a model has the same weight for all locations. Our approach is based on a computer vision method called graph cuts and consists in selecting for each grid point the most appropriate model, while at the same time considering the overall spatial consistency of the resulting field. The performance of the graph cuts approach is assessed based on two experiments: one where the ERA5 reanalyses are considered as the reference, and another involving a perfect model experiment where each model is in turn considered as the reference. We show that the graph cuts approach generally results in lower biases than other model combination approaches such as MMM, while at the same time preserving a similar level of spatial continuity. Supplementary Information: The online version contains supplementary material available at 10.1007/s00382-022-06213-4.

Bone and joint enhancement filtering: Application to proximal femur segmentation from uncalibrated computed tomography datasets.

Methods for reliable femur segmentation enable the execution of quality retrospective studies and building of robust screening tools for bone and joint disease. An enhance-and-segment pipeline is proposed for proximal femur segmentation from computed tomography datasets. The filter is based on a scale-space model of cortical bone with properties including edge localization, invariance to density calibration, rotation invariance, and stability to noise. The filter is integrated with a graph cut segmentation technique guided through user provided sparse labels for rapid segmentation. Analysis is performed on 20 independent femurs. Rater proximal femur segmentation agreement was 0.21 mm (average surface distance), 0.98 (Dice similarity coefficient), and 2.34 mm (Hausdorff distance). Manual segmentation added considerable variability to measured failure load and volume (CVRMS > 5%) but not density. The proposed algorithm considerably improved inter-rater reproducibility for all three outcomes (CVRMS < 0.5%). The algorithm localized the periosteal surface accurately compared to manual segmentation but with a slight bias towards a smaller volume. Hessian-based filtering and graph cut segmentation localizes the periosteal surface of the proximal femur with comparable accuracy and improved precision compared to manual segmentation.