Cut Algorithm Research Articles

Multimodal image fusion based on point-wise mutual information

Multimodal image fusion aims to generate a fused image from different signals that captured by multimodal sensors. Although the images obtained by multimodal sensors have different appearances, the information included in these images might be redundant and noisy. In the previous studies, the fusion rule and their properties that guiding how to merge the features from multiple images is relatively simple functions such as choose-max or weighted average. However, merging the features with redundant information based on these fusion rules may lead to brightness distortion or extra noises, since these fusion rules ignore the spatial consistency of feature selections. In this paper, we propose a novel multimodal image fusion algorithm that focuses on both the transferring the salient structures and maintaining spatial consistency that we consider as fundamental to build our proposed architecture. The proposed algorithm selects features to be transferred into fusion results by a graph cut algorithm, in which the spatial varying smoothness cost is formulated based on the independence between local features measured by point-wise mutual information (PMI). Experiment results demonstrate that, with the straightforward gradient features, the proposed method can obtain state-of-the-art performance on several publicly available multimodal image databases.

Network-Based Coexpression Analysis Identifies Functional and Prognostic Long Noncoding RNAs in Hepatocellular Carcinoma.

Hepatocellular carcinoma (HCC) is a primary liver cancer associated with a growing incidence and extremely high mortality. However, the pathogenic mechanism is still not fully understood. In the present study, we identified 1,631 upregulated and 1,515 downregulated genes and found that cell cycle and metabolism-related pathways or biological processes highly dysregulated in HCC. To assess the biological importance of these DEGs, we carried out weighted gene coexpression network analysis (WGCNA) to identify the functional modules potentially involved in HCC pathogenesis or progression. The five modules were detected with Dynamic Tree Cut algorithm, and GO enrichment analysis revealed that these modules exhibited different biological processes or signaling pathways, such as metabolism-related pathways, cell proliferation-related pathways, and molecules in tumor microenvironment. Moreover, we also observed two immune cells, namely, cytotoxic cells and macrophage enriched in modules grey and brown, respectively, while T helper cell-2 (Th2) was enriched in module turquoise. Among the WGCNA network, four hub long noncoding RNAs (lncRNAs) were identified to be associated with HCC prognostic outcomes, suggesting that coexpression network analysis could uncover lncRNAs with functional importance, which may be associated with prognostic outcomes of HCC patients. In summary, this study demonstrated that network-based analysis could identify some functional modules and some hub-lncRNAs, which may be critical for HCC pathogenesis or progression.

Responsive make‐to‐order supply chain network design

AbstractIn this article, we address two network design problems for a responsive supply chain that consists of make‐to‐order (make‐to‐assemble) facilities facing stochastic demand and service time. The response time to customer orders, critical to the success of the supply chain, is the sum of the flow time in the facility and the delivery time to the customers. The response time performance in our models is measured by the probability that the response time is shorter than a constant. This nonlinear performance measure makes the models less or not tractable. The objective of both problems is to minimize the expected network cost that consists of the cost of delivery to customers as a function of the time of delivery (or the mode of transportation), the fixed cost of locating facilities and the capacity cost as a linear function of the processing capacity. The main decision variables are the number and locations of facilities, the resources allocated to them and the delivery mode selected between facilities and demand. In the first problem, a constraint is imposed to ensure an acceptable response time level. In the second problem, a penalty is charged on the number of days that each unit is delivered later than the targeted response time and is incorporated into the objective function. In the first problem, the probabilistic constraint on the flow time can be linearized and the problem can be formulated as an integer linear programming model. In the second problem, we propose an approximation approach to linearize the objective function and an iterative search and cut algorithm to combine linear approximation with neighborhood search. A multi‐start meta‐heuristic is also suggested. Computational experiments are conducted to evaluate the performance of these solution procedures. Recommendations are made on the basis of the computational results. This appears to be the first study in the area of locating capacitated facilities with stochastic demand to incorporate the delivery mode choice decision and to evaluate the expected congestion cost as a function of the actual flow time in the facilities, instead of the average one. The models developed enable the decision maker to investigate the effect of a combination of facility location, capacity allocation, and delivery mode decisions on the expected network cost and the response time. In addition, the findings of the computational experiments shed light on tuning parameters of approximation algorithms.

Fast Resource Scheduling for Distribution Systems Enabled With Discrete Control Devices

This article proposes a framework for fast short-term scheduling and steady-state voltage control in distribution systems enabled with both continuous control devices, e.g., inverter interfaced distributed generators (DGs) and discrete control devices (DCDs), e.g., on-load tap changers. The voltage-dependent nature of loads is taken into account to further reduce the operating cost by managing the voltage levels. Branch and cut method is applied to handle the integrality constraints associated with the operation of DCDs. A globally convergent trust region algorithm (TRA) is applied to solve the integer-relaxed problems at each node during the branching process. The TRA subproblems are solved using interior point method. To reduce the branching burden of branch and cut algorithm, before applying TRA at each node, a simplified optimization problem is first solved. Using the convergence status and value of objective function of this problem, a faster decision is made on stopping the regarding branch. Solving the simplified problem obviates the application of TRA at most nodes. It is shown that the method converges to the optimal solution with a considerable saving in computation time according to the numerical studies.

Saliency Cuts: Salient Region Extraction based on Local Adaptive Thresholding for Image Information Recognition of the Visually Impaired

In recent years, there has been an increased scope for assistive software and technologies, which help the visually impaired to perceive and recognize natural scene images. In this article, we propose a novel saliency cuts approach using local adaptive thresholding to obtain four regions from a given saliency map. The saliency cuts approach is an effective tool for salient object detection. First, we produce four regions for image segmentation using a saliency map as an input image and applying an automatic threshold operation. Second, the four regions are used to initialize an iterative version of the Grab Cut algorithm and to produce a robust and high-quality binary mask with a full resolution. Lastly, based on the binary mask and extracted salient object, outer boundaries and internal edges are detected by Canny edge detection method. Extensive experiments demonstrate that the proposed method correctly detects and extracts the main contents of the image sequences for delivering visually salient information to the visually impaired people compared to the results of existing salient object segmentation algorithms

Drone Image Segmentation Using Machine and Deep Learning for Mapping Raised Bog Vegetation Communities

The application of drones has recently revolutionised the mapping of wetlands due to their high spatial resolution and the flexibility in capturing images. In this study, the drone imagery was used to map key vegetation communities in an Irish wetland, Clara Bog, for the spring season. The mapping, carried out through image segmentation or semantic segmentation, was performed using machine learning (ML) and deep learning (DL) algorithms. With the aim of identifying the most appropriate, cost-efficient, and accurate segmentation method, multiple ML classifiers and DL models were compared. Random forest (RF) was identified as the best pixel-based ML classifier, which provided good accuracy (≈85%) when used in conjunction graph cut algorithm for image segmentation. Amongst the DL networks, a convolutional neural network (CNN) architecture in a transfer learning framework was utilised. A combination of ResNet50 and SegNet architecture gave the best semantic segmentation results (≈90%). The high accuracy of DL networks was accompanied with significantly larger labelled training dataset, computation time and hardware requirements compared to ML classifiers with slightly lower accuracy. For specific applications such as wetland mapping where networks are required to be trained for each different site, topography, season, and other atmospheric conditions, ML classifiers proved to be a more pragmatic choice.

De-enhancing the dynamic contrast-enhanced breast MRI for robust registration.

Dynamic enhancement causes serious problems for registration of contrast enhanced breast MRI, due to variable uptakes of agent on different tissues or even same tissues in the breast. We present an iterative optimization algorithm to de-enhance the dynamic contrast-enhanced breast MRI and then register them for avoiding the effects of enhancement on image registration. In particular, the spatially varying enhancements are modeled by a Markov Random Field, and estimated by a locally smooth function with boundaries using a graph cut algorithm. The de-enhanced images are then registered by conventional B-spline based registration algorithm. These two steps benefit from each other and are repeated until the results converge. Experimental results show that our two-step registration algorithm performs much better than conventional mutual information based registration algorithm. Also, the effects of tumor shrinking in the conventional registration algorithms can be effectively avoided by our registration algorithm.

A semantic relatedness preserved subset extraction method for language corpora based on pseudo-Boolean optimization

As language corpora have been playing an increasingly important role in the field of Artificial Intelligence (AI) research, lots of extremely large corpora are created. However, a larger corpora size not only increases power and accuracy but also brings redundancy. Therefore, researchers began to emphasize the study of appropriate subset extraction methods. Due to the trade-off between data sufficiency and redundancy, a group of interesting and challenging problems are emerged that are studied in this paper: (1) How to make the resulting subset include as much data as possible under some necessary constraints? (2) How to preserve the potential useful semantic relatedness included in the original corpora while reducing the size of the corpora? For these two problems, existing work mainly focuses on the methods to construct particular subsets for special usage. These methods are limited in their focus. In this paper, we try to address the problems listed above. First, considering the cubic and binary semantic relatedness among tokens, we construct a general system model and formulate the mix problem as a cubic pseudo-Boolean optimization problem. Then, by analyzing the characteristics of the objective function, we transfer the problem into the maximum flow problem of a corresponding graph. Third, we propose a new algorithm by introducing discrete Lagrangian iteration method. We prove that the objective function is supermodular, which allows us to use fast minimum cut algorithms in each iteration step to propose another fast algorithm. Finally, we experimentally validate our new algorithms on several randomly created corpora.

RETRACTED ARTICLE: Design of ensemble classifier using Statistical Gradient and Dynamic Weight LogitBoost for malicious tumor detection

In medical field, the detection of abnormalities in breast is essential to find earlier stage of breast tumor. Conventional semi-supervised ensemble framework based on the normalized cut algorithm developed and it strongly improves the detection accuracy of the resulting images. However, further development of classification algorithm with minimum computational time and cost, several conventional methods limits the classification of tumor. In this paper, Statistical Gradient and Dynamic Weight based LogitBoost (SG-DWL) Ensemble approach is presented to improve the detection rate of malicious tumor. The key objective of SG-DWL Ensemble approach is to increase the malicious tumor performance with higher accuracy and lesser time consumption. The Statistical Gradient Boosting model is introduced as a numerical technique to improve feature set identification in test images. The proposed feature ensemble is formed by concatenating the probability density function, gradient vector and powerful algorithmic framework for feature selection with the best fit feature set is selected. Dynamic Weight based LogitBoost classifier (DW-LC) is applied for malicious tumor detection. This Dynamic Weight based LogitBoost classifier uses Hoeffding tree to achieve high malicious tumor detection rate by reducing the computational complexity involved in the classification of benign and malignant tumor. The performance of the proposed approach is evaluated by comparing it with the existing approaches, and the results improve the classification accuracy with minimum time period for malicious tumor detection.

Superpixel-Based Seamless Image Stitching for UAV Images

Image stitching aims to generate a natural seamless high-resolution panoramic image free of distortions or artifacts as fast as possible. In this article, we propose a new seam cutting strategy based on superpixels for unmanned aerial vehicle (UAV) image stitching. Explicitly, we decompose the issue into three steps: image registration, seam cutting, and image blending. First, we employ adaptive as-natural-as-possible (AANAP) warps for registration, obtaining two aligned images in the same coordinate system. Then, we propose a novel superpixel-based energy function that integrates color difference, gradient difference, and texture complexity information to search a perceptually optimal seam located in continuous areas with high similarity. We apply the graph cut algorithm to solve the problem and thereby conceal artifacts in the overlapping area. Finally, we utilize a superpixel-based color blending approach to eliminate visible seams and achieve natural color transitions. Experimental results demonstrate that our method can effectively and efficiently realize seamless stitching, and is superior to several state-of-the-art methods in UAV image stitching.

Skin disease detection and segmentation using dynamic graph cut algorithm and classification through Naive Bayes classifier

Abstract The largest organ and the outer covering of the human body is the skin. With seven layers of it covering the other organs inside, skin is one of the important part to take care of. A skin condition is one which affects the integumentary system and that includes a wide variety of diseases including dermatoses. Classifications of these skin conditions are always a challenge for any medical practitioner and they look at the machine learning systems to assist them in predicting and classifying the skin conditions. This in turn will help to cure or at least reduce the effect. If the skin symptoms such as acne, cellulitis, candidiasis, varicella, scleroderma, fungal skin, psoriasis, inflamed skin condition, etc. are left without treatment in its initial stage, then they can effect in different health impediments leading to even death. Image partitioning is a method which supports with the skin disease detection. Any abnormal skin growth is referred to as skin lesion which could either be primary or secondary. Graph cut algorithms are debated and used in the literature for variety of purposes including image smoothing, image segmentation and other problems involving energy minimization as objective. In this work, we intend to use a novel dynamic graph cut algorithm for skin lesion segmentation followed by a probabilistic classifier called as Naive Bayes classifier for skin disease classification purposes. We have used ISIC 2017 dataset for testing our proposed method and found that the results outperform many state of the art methods including FCN and SegNet by 6.5% and 8.7% respectively. This dataset is available at the International Skin Imaging Collaboration (ISIC) website for public study and experimentation. In terms of accuracy, we could achieve 94.3% for benign cases, 91.2% for melanoma and 92.9% for keratosis on this data set.

Classification of brain tumours using radiomic features on MRI

Glioma is one of the most common brain tumours among the diagnoses of existing brain tumours. Glioma grades are important factors that should be known in the treatment of brain tumours. In this study, the radiomic features of gliomas were analysed and glioma grades were classified by Gaussian Naive Bayes algorithm. Glioma tumours of 121 patients of Grade II and Grade III were examined. The glioma tumours were segmented with the Grow Cut Algorithm and the 3D feature of tumour magnetic resonance imaging images were obtained with the 3D Slicer programme. The obtained quantitative values were statistically analysed with Spearman and Mann–Whitney U tests and 21 features with statistically significant properties were selected from 107 features. The results showed that the best performing among the algorithms was Gaussian Naive Bayes algorithm with 80% accuracy. Machine learning and feature selection techniques can be used in the analysis of gliomas as well as pathological evaluations in glioma grading processes.
 
 Keywords: Radiomics, glioma, naive bayes.

Analysis of Cell Death by Image Processing

Abstract In this paper, we present a graph theoretical approach to image processing with focus on the analysis of the biological data. We use the graph cut algorithms and extend them to obtain a segmentation of the biological cells. We introduce an utterly new algorithm for analysis of the resulting data and for sorting them into three main categories, which correspond to the biological cell death, based on the mathematical properties of the segmented elements.

Linear edge costs and labeling algorithms: The case of the time‐dependent vehicle routing problem with time windows

AbstractIn this paper we implement a branch‐price and cut algorithm for a time dependent vehicle routing problem with time windows in which the goal is to minimize the total route duration. The travel time between two customers is given by a piecewise linear function on the departure time and, thus, it need not remain fixed along the planning horizon. We discuss different alternatives for the implementation of these linear functions within the labeling algorithm applied to solve the pricing problem. We also provide a tailored implementation for one of these alternatives, relying on efficient data structures for storing the labels, and show several strategies to accelerate the algorithm. Computational results show that the proposed techniques are effective and improve the column generation step, solving all instances with 25 customers, 49 of 56 with 50 customers, and many instances with 100 customers. Furthermore, heuristic adaptations are able to find good quality solutions in reasonable computation times.

Higher-order potentials for video object segmentation in bilateral space

We propose an effective approach to make segmentation for objects in videos with an initial input of the object masks in a few frames of the source video. In this method, we cast the segmentation task as a Markov Random Field (MRF) labeling problem. Different from the conventional MRF models, our model uses an additional term of higher-order potential to better propagate the global consistency among frames. The higher-order potential presented in this paper is significant for the proposed method because of its capability to keep the long-range consistency during segmentation. In order to make the MRF energy minimized, we also introduce a smart skill that makes the intractable higher-order potential “invisible” during the optimization so that the problem can be solved simply by applying a standard graph cut algorithm. Besides, the entire process is operated in a bilateral space, where the labeling can be inferred efficiently on the vertices that are sampled regularly from the bilateral grid. The results of a comparison of our method with a number of recently developed methods show that it performs favorably against state-of-the-art algorithms on multiple benchmark data sets in view of accuracy and achieves a much faster runtime performance.

Discriminative Training of Conditional Random Fields with Probably Submodular Constraints

Problems of segmentation, denoising, registration and 3D reconstruction are often addressed with the graph cut algorithm. However, solving an unconstrained graph cut problem is NP-hard. For tractable optimization, pairwise potentials have to fulfill the submodularity inequality. In our learning paradigm, pairwise potentials are created as the dot product of a learned vector w with positive feature vectors. In order to constrain such a model to remain tractable, previous approaches have enforced the weight vector to be positive for pairwise potentials in which the labels differ, and set pairwise potentials to zero in the case that the label remains the same. Such constraints are sufficient to guarantee that the resulting pairwise potentials satisfy the submodularity inequality. However, we show that such an approach unnecessarily restricts the capacity of the learned models. Guaranteeing submodularity for all possible inputs, no matter how improbable, reduces inference error to effectively zero, but increases model error. In contrast, we relax the requirement of guaranteed submodularity to solutions that are probably approximately submodular. We show that the conceptually simple strategy of enforcing submodularity on the training examples guarantees with low sample complexity that test images will also yield submodular pairwise potentials. Results are presented in the binary and muticlass settings, showing substantial improvement from the resulting increased model capacity.

Hole-filling approach based on convolutional neural network for depth image-based rendering view synthesis

Depth image-based rendering (DIBR) has recently received significant attention as an efficient concept for generating virtual views. One of the main challenges in the view synthesis process is that holes will appear in the background (BG) of the virtual views because of occlusion. The existing methods either utilize spatial or temporal information from the original views and depth maps to fill the holes. However, the performance of the existing methods is seriously affected by the quality of the depth maps. Foreground (FG) texture will be wrongly introduced to the BG when the depth maps are not accurate enough. We develop a hole-filling approach for DIBR based on convolutional neural network, which shows reduced dependency on the quality of the depth map. More specifically, FG objects are detected on the RGB image by combining the Laplacian operator with a graph cut algorithm to avoid dependencies on the depth map quality. Furthermore, we learn an end-to-end mapping between the warped virtual view and the ground-truth view image with FG awareness. The learning-based approach exhibits excellent superiority in terms of saving time. The experimental results also show that the proposed approach outperforms current state-of-the-art techniques quantitatively and qualitatively.

A novel method to model hepatic vascular network using vessel segmentation, thinning, and completion.

The accurate modeling of the liver vessel network structure is an important prerequisite for developing a preoperative plan for the liver. Considering that extracting liver blood vessels from patient's abdominal computed tomography(CT) images requires several manual operations, this study proposed an automatic segmentation method of liver vessels based on graph cut, thinning, and vascular combination, which can obtain a complete liver vascular network. First, the CT image was preprocessed by grayscale mapping based on sigmoid function, vessel enhancement based on Hessian filter, and denoising based on anisotropic filter to enhance the grayscale contrast between the vascular and non-vascular parts of the liver. Then, the liver vessels were initially segmented based on the improved three-dimensional graph cut algorithm. Based on the obtained liver vascular structure, the vessel centerline of the liver was then extracted by the proposed thinning algorithm that continuously traversed the foreground voxel points and iteratively deleted the simple points. Finally, the combination of vascular centerline optimization was used to predict and link the vascular centerline fractured portion. The under-segmented liver vessels were complemented based on the complete vascular centerline tree. To verify the proposed hepatic vascular segmentation and complementation algorithm, the open 3D Image Reconstruction for Comparison of Algorithm Database (3Dircadb) was applied to test and quantify the results. The results showed that the proposed algorithm can accurately and effectively segment the vascular network structure from abdominal CT images, and the proposed vascular complementation method can restore the true information of under-segmented liver vessels. Graphical abstract A novel hepatic vessel segmentation method from abdominal CT images was proposed, including graph cut algorithm, centerline extraction, and broken vessel completion. First, the graph cut algorithm was used to obtain the initial segmentation result. Then, the centerline of the initial segmentation result was extracted. Finally, the initial segmentation result was optimized through centerline analysis.

Semi-automatic mapping of dyke and dyke-related fractures using UAV-based photogrammetric data: A case study from Sijiao Island, coastal Southeastern China

Mapping dyke and dyke-related fractures is important for magma propagation and emplacement mechanisms analysis. Here, we propose a semi-automatic workflow for dyke and fracture mapping using unmanned aerial vehicle (UAV)-based photogrammetric data. Our workflow integrates mean-shift and GrabCut in one cut (GCOC) algorithms to segment dykes, and harnesses the phase symmetry algorithm and a series of follow-up actions to detect fractures. Sijiao Island of Zhejiang Province (SE China) is examined as a case study to validate the approaches. With our workflow, the generated dyke map attains more than a 97% recall rate when compared with the results of expert interpretation, and the ratio of the number of detected false fracture segments to the number of total segments is less than 18%. The dyke strike and fracture strike calculation match well with field measurements. Moreover, by combining visual interpretation and the resulting geometric information, our results manifest that dyke magma invaded along pre-existing fractures to form dykes on Sijiao Island, and the difference between dyke fractures and host rock fractures may be because of the effect of major changes of regional stress. This workflow has great potential for high-resolution terrestrial mapping.

Outfit Selection Using Colour Matching Application and Image Processing

The application is a fashion clothing app, that would assist users with their daily or occasional clothing choices, by pairing various outfits that have been extracted using the grab cut algorithm. These outfits will be placed over a 2D model that gives a visual result for the user to make a better choice. The algorithm for foreground extraction in this project is the Grab Cut Algorithm. The 2D model contains a skin tone feature that can be customized easily. A secure login page protects the user data. A ’Share’ feature which allows users to share their customized ideas on a 2D model to other users on various social media platforms. Another feature ’hub’ where the model designs can be shared and rated by different users.