Iterative Clustering Algorithm Research Articles

Semantic Recognition of Human-Object Interactions via Gaussian-Based Elliptical Modeling and Pixel-Level Labeling

Human-Object Interaction (HOI) recognition, due to its significance in many computer vision-based applications, requires in-depth and meaningful details from image sequences. Incorporating semantics in scene understanding has led to a deep understanding of human-centric actions. Therefore, in this research work, we propose a semantic HOI recognition system based on multi-vision sensors. In the proposed system, the de-noised RGB and depth images, via Bilateral Filtering (BLF), are segmented into multiple clusters using a Simple Linear Iterative Clustering (SLIC) algorithm. The skeleton is then extracted from segmented RGB and depth images via Euclidean Distance Transform (EDT). Human joints, extracted from the skeleton, provide the annotations for accurate pixel-level labeling. An elliptical human model is then generated via a Gaussian Mixture Model (GMM). A Conditional Random Field (CRF) model is trained to allocate a specific label to each pixel of different human body parts and an interaction object. Two semantic feature types that are extracted from each labeled body part of the human and labelled objects are: Fiducial points and 3D point cloud. Features descriptors are quantized using Fisher's Linear Discriminant Analysis (FLDA) and classified using K-ary Tree Hashing (KATH). In experimentation phase the recognition accuracy achieved with the Sports dataset is 92.88%, with the Sun Yat-Sen University (SYSU) 3D HOI dataset is 93.5% and with the Nanyang Technological University (NTU) RGB+D dataset it is 94.16%. The proposed system is validated via extensive experimentation and should be applicable to many computer-vision based applications such as healthcare monitoring, security systems and assisted living etc.

A Novel Method for Automated Lung Region Segmentation in Chest X-Ray Images

Detecting and segmenting the lung regions in chest X-ray images is an important part in artificial intelligence-based computer-aided diagnosis/detection (AI-CAD) systems for chest radiography. However, if the chest X-ray images themselves are used as training data for the AI-CAD system, the system might learn the irrelevant image-based information resulting in the decrease of system’s performance. In this study, we propose a lung region segmentation method that can automatically remove the shoulder and scapula regions, mediastinum, and diaphragm regions in advance from various chest X-ray images to be used as learning data. The proposed method consists of three main steps. First, employ the simple linear iterative clustering algorithm, the lazy snapping technique and local entropy filter to generate an entropy map. Second, apply morphological operations to the entropy map to obtain a lung mask. Third, perform automated segmentation of the lung field using the obtained mask. A total of 30 images were used for the experiments. In order to verify the effectiveness of the proposed method, two other texture maps, namely, the maps created from the standard deviation filtering and the range filtering, were used for comparison. As a result, the proposed method using the entropy map was able to appropriately remove the unnecessary regions. In addition, this method was able to remove the markers present in the image, but the other two methods could not. The experimental results have revealed that our proposed method is a highly generalizable and useful algorithm. We believe that this method might act an important role to enhance the performance of AI-CAD systems for chest X-ray images.

Salient Object Detection Based on Multiscale Segmentation and Fuzzy Broad Learning

Abstract Saliency detection has been a hot topic in the field of computer vision. In this paper, we propose a novel approach that is based on multiscale segmentation and fuzzy broad learning. The core idea of our method is to segment the image into different scales, and then the extracted features are fed to the fuzzy broad learning system (FBLS) for training. More specifically, it first segments the image into superpixel blocks at different scales based on the simple linear iterative clustering algorithm. Then, it uses the local binary pattern algorithm to extract texture features and computes the average color information for each superpixel of these segmentation images. These extracted features are then fed to the FBLS to obtain multiscale saliency maps. After that, it fuses these saliency maps into an initial saliency map and uses the label propagation algorithm to further optimize it, obtaining the final saliency map. We have conducted experiments based on several benchmark datasets. The results show that our solution can outperform several existing algorithms. Particularly, our method is significantly faster than most of deep learning-based saliency detection algorithms, in terms of training and inferring time.

Imaging-based characterization of convective tissue properties

Convective transport is an important phenomenon for nanomedicine delivery. We present an imaging-based approach to recover tissue properties that are significant in the accumulation of nanoparticles delivered via systemic methods. The classical pharmacokinetic analysis develops governing equations for the particle transport from a first principle mass balance. Fundamentally, the governing equations for compartmental mass balance represent a spatially invariant mass transport between compartments and do not capture spatially variant convection phenomena. Further, the parameters recovered from this approach do not necessarily have direct meaning with respect to the governing equations for convective transport. In our approach, a framework is presented for directly measuring permeability in the sense of Darcy flow through porous tissue. Measurements from our approach are compared to an extended Tofts model as a control. We demonstrate that a pixel-wise iterative clustering algorithm may be applied to reduce the parameter space of the measurements. We show that measurements obtained from our approach are correlated with measurements obtained from the extended Tofts model control. These correlations demonstrate that the proposed approach contains similar information to an established compartmental model and may be useful in providing an alternative theoretical framework for parameterizing mathematical models for treatment planning and diagnostic studies involving nanomedicine where convection dominated effects are important.

Super‐pixel image segmentation algorithm based on adaptive equalisation feature parameters

Image segmentation is a key step in the process of image data processing. The quality of image segmentation will directly affect the accuracy of image cognitive understanding. The purpose of image segmentation is to divide the image into regions with specific semantics. For the simple linear iterative clustering (SLIC) algorithm, the feature equalisation parameters need to be set manually during image segmentation, which results in the lack of segmentation effects and slow processing time. By introducing the theory of intermediary mathematics, an improved adaptive SLIC super-pixel algorithm is proposed, which can adaptive generate characteristic equalisation parameters according to the specific situation of the image, thereby simplifying the operation steps and improving the image segmentation effect. After experimental verification and analysis, compared with the original SLIC algorithm and several other super-pixel contrast algorithms, the algorithm in this study can effectively shorten the processing time and achieve a better segmentation effect.

A clustering detector with graph theory for blind detection of spatial modulation systems

This paper considers the blind detection of spatial modulation systems multiple-input multiple-output systems. In this system, we propose a clustering detection framework with graph theory that conducts signal detection without the training of channel state information. In detail, firstly, by dynamically controlling the size of each cluster, we transform the original optimization problem of the traditional K-means clustering detector into a new optimization problem. In addition, the cluster assignment subproblem of the iterative clustering algorithm for solving the new optimization problem makes it equivalent to a minimum cost flow linear network optimization problem of graph theory, which can be addressed by the breadth-first algorithm. Moreover, a novel clustering detector with the breadth-first algorithm is presented correspondingly. Numerical results show that the proposed detector is efficient in avoiding the undesired local optima and can closely approach the performance of the optimal detector.

Wood Defect Detection Based on Depth Extreme Learning Machine

The deep learning feature extraction method and extreme learning machine (ELM) classification method are combined to establish a depth extreme learning machine model for wood image defect detection. The convolution neural network (CNN) algorithm alone tends to provide inaccurate defect locations, incomplete defect contour and boundary information, and inaccurate recognition of defect types. The nonsubsampled shearlet transform (NSST) is used here to preprocess the wood images, which reduces the complexity and computation of the image processing. CNN is then applied to manage the deep algorithm design of the wood images. The simple linear iterative clustering algorithm is used to improve the initial model; the obtained image features are used as ELM classification inputs. ELM has faster training speed and stronger generalization ability than other similar neural networks, but the random selection of input weights and thresholds degrades the classification accuracy. A genetic algorithm is used here to optimize the initial parameters of the ELM to stabilize the network classification performance. The depth extreme learning machine can extract high-level abstract information from the data, does not require iterative adjustment of the network weights, has high calculation efficiency, and allows CNN to effectively extract the wood defect contour. The distributed input data feature is automatically expressed in layer form by deep learning pre-training. The wood defect recognition accuracy reached 96.72% in a test time of only 187 ms.

Automated Superpixel-borders-guided Deformation Image Registration for Adaptive Radiotherapy

Registration between the baseline and follow-up lung computed tomography (CT) volumes plays an important role in computer-aided diagnosis and following-up care during adaptive radiotherapy. Diffeomorphic log-Demons as state of the art in Demons implementations is restricted to relatively small deformations, low accuracy, and ignoring some prior structural features. In this paper, an automated superpixel-borders-guided deformation image registration (SBG-DIR) algorithm is proposed. The proposed SBG-DIR method uses the simple linear iterative clustering (SLIC) algorithm to automated superpixels generation. Incorporation of superpixel borders into registration algorithm is implemented by a new similarity criterion based on the binary volume representation of superpixel borders. The binary volume representation enables accurate preserving motion boundaries, contributes to a faster convergence of the objective function and eliminates errors caused by manual interaction. In addition, a subtraction volume is produced by the intensity difference between the first time point CT volume and its warped follow-up CT volume. The subtraction volume can be used for detection of tumor tissue growth or shrinkage, which is an essential part of a CT-based diagnosis. Moreover, to ensure the topology preservation of biological objects, our proposed SBG-DIR method is implemented in the space of diffeomorphisms, in which meaningful biological shapes can be found. Compared with the state-of-the-art Demons, the proposed SBG-DIR method doesn’t require any additional optimization, yields a faster convergence and is more accurate and efficient in recovering large deformations. Experimental results indicate that the proposed SBG-DIR method performed better than the state-of-the-art Demons algorithms.

Hybrid method combining superpixel, supervised learning, and random walk for glioma segmentation

AbstractCurrently, the analysis of magnetic resonance imaging (MRI) brain images of pathological patients is performed manually, both for the recognition of brain structures or lesions and for their characterization. Physicians sometimes encounter difficulties in interpreting these images for a reliable diagnosis of the patient's condition. This is due to the difficulty of detecting the nature of the lesions, particularly glioma. Glioma is one of the most common tumors, and one of the most difficult to detect because of its shape, irregularities, and ambiguous limits. The segmentation of these tumors is one of the most crucial steps for their classification and surgical planning. This article presents a new, accurate, and automatic approach for the precise segmentation of early gliomas (benign tumors), combining the random walk (RW) algorithm and the simple linear iterative clustering algorithm. The study was carried out in four steps. The first step consisted of decomposing the image into superpixels to obtain an initial outline of the tumor. The superpixels were generated using the SLIC algorithm. In the second step, for each superpixel, a set of statistical and multifractal characteristics were calculated (gray‐level co‐occurrence matrix, multifractal detrending moving average). In the third step, the superpixels were classified using a supervised random forest (RF) type classier into healthy or tumorous brain tissue. In the final step, the contour of the detected tumor was enhanced using the customized RW algorithm. The proposed method was evaluated using the Brain Tumor Image Segmentation Challenge 2013 database. The results obtained are competitive compared to other existing methods.

Discriminating crops/weeds in an upland rice field from UAV images with the SLIC-RF algorithm

ABSTRACT In this study, we propose a method for discriminating crops/weeds in upland rice fields using a commercial unmanned aerial vehicles (UAVs) and red-green-blue (RGB) cameras with the simple linear iterative clustering (SLIC) algorithm and random forest (RF) classifier. In the SLIC-RF algorithm, we evaluated different combinations of input features: three color spaces (RGB, hue-saturation-brightness [HSV], CIE-L*a*b), canopy height model (CHM), spatial texture (Texture) and four vegetation indices (VIs) (excess green [ExG], excess red [ExR], green-red vegetation index [GRVI] and color index of vegetation extraction [CIVE]). Among the color spaces, the HSV-based SLIC-RF model showed the best performance with the highest out-of-bag (OOB) accuracy (0.904). The classification accuracy was improved by the combination of HSV with CHM, Texture, ExG, or CIVE. The highest OOB accuracy (0.915) was obtained from the HSV+Texture combination. The greatest errors from the confusion matrix occurred in the classification between crops and weeds, while soil could be classified with a very high accuracy. These results suggest that with the SLIC-RF algorithm developed in this study, rice and weeds can be discriminated by consumer-grade UAV images with acceptable accuracy to meet the needs of site-specific weed management (SSWM) even in the early growth stages of small rice plants..

Penalty term based suitable fuzzy intuitionistic possibilistic clustering: analyzing high dimensional gene expression cancer database

The aim of this paper is to identify the co-expressed potential genes that may serve for the development of the portions of normal or tumor. This paper differentiates the co-expressed genes into normal samples and tumor samples from gene expression dataset GSE25066. Since the dataset has vague boundaries and having common characteristics between the clusters, identifying the subgroups contain similar gene expression is really a tricky task one. Therefore, this paper introduces an effective fuzzy iterative clustering algorithm by incorporating kernel function, possibilistic c-means, fuzzy memberships, neighborhood information, median of neighboring objects and penalty term. The performances of the proposed clustering techniques have been shown through the succession experimental works on GSE25066. The effects of clustering results have been proved through comparing the resulted classes with ground truth.

Pattern Classification Based on RBF Networks with Self-Constructing Clustering and Hybrid Learning

Radial basis function (RBF) networks are widely adopted to solve problems in the field of pattern classification. However, in the construction phase of such networks, there are several issues encountered, such as the determination of the number of nodes in the hidden layer, the form and initialization of the basis functions, and the learning of the parameters involved in the networks. In this paper, we present a novel approach for constructing RBF networks for pattern classification problems. An iterative self-constructing clustering algorithm is used to produce a desired number of clusters from the training data. Accordingly, the number of nodes in the hidden layer is determined. Basis functions are then formed, and their centers and deviations are initialized to be the centers and deviations of the corresponding clusters. Then, the parameters of the network are refined with a hybrid learning strategy, involving hyperbolic tangent sigmoid functions, steepest descent backpropagation, and least squares method. As a result, optimized RBF networks are obtained. With this approach, the number of nodes in the hidden layer is determined and basis functions are derived automatically, and higher classification rates can be achieved. Furthermore, the approach is applicable to construct RBF networks for solving both single-label and multi-label pattern classification problems.

AN UNSUPERVISED LABELING APPROACH FOR HYPERSPECTRAL IMAGE CLASSIFICATION

Abstract. The application of hyperspectral image analysis for land cover classification is mainly executed in presence of manually labeled data. The ground truth represents the distribution of the actual classes and it is mostly derived from field recorded information. Its manual generation is ineffective, tedious and very time-consuming. The continuously increasing amount of proprietary and publicly available datasets makes it imperative to reduce these related costs. In addition, adequately equipped computer systems are more capable of identifying patterns and neighbourhood relationships than a human operator. Based on these facts, an unsupervised labeling approach is presented to automatically generate labeled images used during the training of a convolutional neural network (CNN) classifier. The proposed method begins with the segmentation stage where an adapted version of the simple linear iterative clustering (SLIC) algorithm for dealing with hyperspectral data is used. Consequently, the Hierarchical Agglomerative Clustering (HAC) and Fuzzy C-Means (FCM) algorithms are employed to efficiently group similar superpixels considering distances with respect to each other. The distinct utilization of these clustering techniques defines a complementary stage for overcoming class overlapping during image generation. Ultimately, a CNN classifier is trained using the computed image to pixel-wise predict classes on unseen datasets. The labeling results, obtained using two hyperspectral benchmark datasets, indicate that the current approach is able to detect objects boundaries, automatically assign class labels to the entire dataset and to classify new data with a prediction certainty of 90%. Additionally, this method is also capable of achieving better classification accuracy and visual correspondence with reality than the ground truth images.

A Primer on Machine Learning.

A Primer on Machine Learning.

Superpixel-based segmentation algorithm for mature citrus.

With the decrease of agricultural labors and the increase in production costs, harvesting robots have become a research hotspot in recent years. To guide harvesting robots to pick mature citrus more precisely under variable illumination conditions, an image segmentation algorithm based on superpixel was proposed. Efficient simple linear iterative clustering (SLIC) algorithm which takes similarity of adjacent pixels into account was adopted to segment the images captured under variable illumination conditions into superpixels. The color and texture features of these superpixels were extracted and fused into feature vectors as descriptors to train backpropagation neural networks (BPNN) classifier in the next step. The adjacency information of superpixels was considered by calculating the global-local binary pattern (LBP) in R component images when extracting texture features. To accelerate the classification process, the mean of Cr-Cb image was utilized to find superpixels of interest which were regarded as candidates of citrus superpixels. These candidates were then classified by a pre-trained BPNN model with superpixel-level accuracy of 98.77% and pixel-level accuracy of 94.96%, while the average time to segment one image was 0.4778 s. Therefore, the results indicated that a superpixel-based segmentation algorithm toward citrus images had decent light robustness as well as high accuracy that could guide harvesting robot to pick mature citrus efficiently. Keywords: superpixel, image segmentation, BPNN, variable illumination, mature citrus DOI: 10.25165/j.ijabe.20201304.5607 Citation: Yang Q H, Chen Y Q, Xun Y, Bao G J. Superpixel-based segmentation algorithm for mature citrus. Int J Agric & Biol Eng, 2020; 13(4): 166–171.

EARTHQUAKE-DAMAGED REGIONS DETECTION FROM HIGH RESOLUTION IMAGE BASED ON SUPER-PIXEL SEGMENTATION AND DEEP LEARNING

Abstract. Accurate detection and automatic processing of earthquake-damaged regions is essential for effective rescue and post-disaster reconstruction. In this study, we proposed a Combined Super-pixel Segmentation and AlexNet Detection approach (CSSAD) for automatically extracting damaged regions from post-earthquake high-resolution images. Simple Linear Iterative Clustering (SLIC) algorithm was used to segment the high resolution images to obtain more homogeneous geo-objects. Multiscale samples database, which took the different scale effect of damaged regions into account, was constructed based on the geometric centre of each super-pixel. AlexNet, which achieved the automatic extraction of high-level features and accurate identification of target geo-objects, was used to detect the damaged regions. To enhance the localization accuracy, the output of AlexNet was further refined using super-pixel segmentations and masked out of shadow and vegetation. Compared with traditional method, the proposed approach effectively reduces the false and missed detection ratio at least 10 percent.

Cassava NDVI Analysis: A Nonlinear Mixed Model Approach Based on UAV-Imagery

Imagery data captured by unmanned aerial vehicles ( $${\text{UAVs}}$$ ) have become important tools to study crop growth and development in experimental agronomic research. Moreover, growth curve analysis and nonlinear mixed-effects models ( $${\text{NLME}}$$ ) are increasingly being used to study nonlinear crop responses in the context of repeated measurements. An $${\text{NLME}}$$ to fit the Normalized Difference Vegetation Index ( $${\text{NDVI}}$$ ) based on over-segmented $${\text{UAV}}$$ -imagery of a cassava experimental field is presented. To study the parameters’ variability and error propagation, a resampling analysis and posterior $${\text{NLME}}$$ fit for different sample sizes were performed. High-resolution multispectral images (7 cm resolution) were captured in a cassava experimental field sown under optimal production conditions. $${\text{NDVI}}$$ for individual plants was dynamically calculated by performing a supervised classification based on over-segmented remote images using the simple linear iterative clustering algorithm for near infra-red–red–green bands. A three-parameter logistic function was adopted as a growth curve in function of growing degree days ( $$^\circ {\text{C}}$$ ) and fitted with an $${\text{NLME}}$$ approach. Inherent cassava $${\text{NDVI}}$$ variability due to pixel discretization, plant architecture and cassava growth was found. On average, $${\text{NDVI}}$$ values ranged from 0.35 to 0.40 for first development stages (295, 340 and 372 $$^\circ {\text{C }}$$ days) until 0.77 for maturity (1606 $$^\circ {\text{C }}$$ days). $${\text{NDVI}}$$ plant variability was correctly addressed considering the three parameters as random effects showing small root-mean-square error. Resampling analysis proved that a suitable accuracy parameter estimation can be performed with fewer individuals (plants) which might represent an agricultural experimental design with fewer experimental units. While the application of $${\text{UAV}}$$ imaging and nonlinear mixed models in determining $${\text{NDVI}}$$ curves is a relevant methodological framework, further research towards the optimization of experiment sample size is required.

An optimized iterative clustering framework for recognizing speech

In the recent years, many research methodologies are proposed to recognize the spoken language and translate them to text. In this paper, we propose a novel iterative clustering algorithm that makes use of the translated text and reduces error in it. The proposed methodology involves three steps executed over many iterations, namely: (1) unknown word probability assignment, (2) multi-probability normalization, and (3) probability filtering. In the first case, each iteration learns the unknown words from previous iterations and assigns a new probability to the unknown words based on the temporary results obtained in the previous iteration. This process continues until there are no unknown words left. The second case involves normalization of multiple probabilities assigned to a single word by considering neighbour word probabilities. The last step is to eliminate probabilities below the threshold, which ensures the reduction of noise. We measure the quality of clustering with many real-world benchmark datasets. Results show that our optimized algorithm produces more accurate clustering compared to other clustering algorithms.

Phase I non-linear profiles monitoring using a modified Hausdorff distance algorithm and clustering analysis

PurposeThe purpose of this paper is to propose a new non-parametric phase I control chart for the problem of non-linear profile outlier detection.Design/methodology/approachThe proposed non-parametric method is based on a modified Hausdorff distance, which does not require a restrictive assumption on the form of profiles. By obtaining the distance between each profile and the baseline profile, the authors introduced an iterative optimization clustering algorithm to identify outliers by clustering distances.FindingsThe simulation results show that the proposed method can distinguish outliers for structural changes of non-linear profiles. The authors also present a real industrial case example to highlight how practitioners can implement and make use of the proposed control chart in outlier detection applications, and it achieves higher accuracy in the outlier detection of complex profiles.Practical implicationsThe research results of this paper can be applied to any manufacturing or service system whose quality characteristics are characterized by non-linear profiles. This new approach provides quality practitioners a better decision-making tool for non-linear profile outlier detection.Originality/valueDue to the complexity of real-world applications, the non-linear profiles monitoring problem is yet to be addressed. However, the related research still remains rare. And the authors’ proposed non-linear profile control chart, which does not require a restrictive assumption on the form of profiles, shows its applicability and superiority in simulation study and real-world case.

Deep learning-based unsupervised representation clustering methodology for automatic nuclear reactor operating transient identification

Transient identification of condition monitoring data in nuclear reactor is important for system health assessment. Conventionally, the operating transients are correlated with the pre-designed ones by human operators during operations. However, due to necessary conservatism and significant differences between the operating and pre-designed transients, it has been less effective to manually identify transients, that usually contribute to different system degradation modes. This paper proposes a deep learning-based unsupervised representation clustering method for automatic transient pattern recognition based on the on-site condition monitoring data. Sample entropy is used as indicator for transient extraction, and a pre-training stage is implemented using an auto-encoder architecture for learning high-level features. An iterative representation clustering algorithm is further proposed to enhance the clustering effects, where a novel distance metric learning strategy is integrated. Experiments on a real-world nuclear reactor condition monitoring dataset validate the effectiveness and superiority of the proposed method, which provides a promising tool for transient identification in the real industrial scenarios. This study offers a new perspective in exploring unlabeled data with deep learning, and the end-to-end implementation scheme facilitates applications in the real nuclear industry.