Algebraic Multi-Layer Network: Key Concepts.

The paper refers to interdisciplinary research in the areas of hierarchical cluster analysis of big data and ordering of primary data to detect objects in a color or in a grayscale image. To perform this on a limited domain of multidimensional data, an NP-hard problem of calculation of close to optimal piecewise constant data approximations with the smallest possible standard deviations or total squared errors (approximation errors) is solved. The solution is achieved by revisiting, modernizing, and combining classical Ward's clustering, split/merge, and K-means methods. The concepts of objects, images, and their elements (superpixels) are formalized as structures that are distinguishable from each other. The results of structuring and ordering the image data are presented to the user in two ways, as tabulated approximations of the image showing the available object hierarchies. For not only theoretical reasoning, but also for practical implementation, reversible calculations with pixel sets are performed easily, as with individual pixels in terms of Sleator-Tarjan Dynamic trees and cyclic graphs forming an Algebraic Multi-Layer Network (AMN). The detailing of the latter significantly distinguishes this paper from our prior works. The establishment of the invariance of detected objects with respect to changing the context of the image and its transformation into grayscale is also new.

Pixel-Level Clustering of Hematoxylin-Eosin-Stained Sections of Mouse and Human Biliary Tract Cancer.

We previously established mouse models of biliary tract cancer (BTC) based on the injection of cells with biliary epithelial stem cell properties derived from KRAS(G12V)-expressing organoids into syngeneic mice. The resulting mouse tumors appeared to recapitulate the pathological features of human BTC. Here we analyzed images of hematoxylin and eosin (H&E) staining for both the mouse tumor tissue and human cholangiocarcinoma tissue by pixel-level clustering with machine learning. A pixel-clustering model that was established via training with mouse images revealed homologies of tissue structure between the mouse and human tumors, suggesting similarities in tumor characteristics independent of animal species. Analysis of the human cholangiocarcinoma tissue samples with the model also revealed that the entropy distribution of cancer regions was higher than that of noncancer regions, with the entropy of pixels thus allowing discrimination between these two types of regions. Histograms of entropy tended to be broader for noncancer regions of late-stage human cholangiocarcinoma. These analyses indicate that our mouse BTC models are appropriate for investigation of BTC carcinogenesis and may support the development of new therapeutic strategies. In addition, our pixel-level clustering model is highly versatile and may contribute to the development of a new BTC diagnostic tool.

Extraction of bouton-like structures from neuropil calcium imaging data.

The neuropil, the plexus of axons and dendrites, plays a critical role in operating the circuit processing of the nervous system. Revealing the spatiotemporal activity pattern within the neuropil would clarify how the information flows throughout the nervous system. However, calcium imaging to examine the circuit dynamics has mainly focused on the soma population due to their discrete distribution. The development of a methodology to analyze the calcium imaging data of a densely packed neuropil would provide us with new insights into the circuit dynamics. Here, we propose a new method to decompose calcium imaging data of the neuropil into populations of bouton-like synaptic structures with a standard desktop computer. To extract bouton-like structures from calcium imaging data, we introduced a new type of modularity, a widely used quality measure in graph theory, and optimized the clustering configuration by a simulated annealing algorithm, which is established in statistical physics. To assess this method's performance, we conducted calcium imaging of the neuropil of Drosophila larvae. Based on the obtained data, we established artificial neuropil imaging datasets. We applied the decomposition procedure to the artificial and experimental calcium imaging data and extracted individual bouton-like structures successfully. Based on the extracted spatiotemporal data, we analyzed the network structure of the central nervous system of fly larvae and found it was scale-free. These results demonstrate that neuropil calcium imaging and its decomposition could provide new insight into our understanding of neural processing.

A Model of Pixel and Superpixel Clustering for Object Detection.

The paper presents a model of structured objects in a grayscale or color image, described by means of optimal piecewise constant image approximations, which are characterized by the minimum possible approximation errors for a given number of pixel clusters, where the approximation error means the total squared error. An ambiguous image is described as a non-hierarchical structure but is represented as an ordered superposition of object hierarchies, each containing at least one optimal approximation in g0 = 1, 2,..., etc., colors. For the selected hierarchy of pixel clusters, the objects-of-interest are detected as the pixel clusters of optimal approximations, or as their parts, or unions. The paper develops the known idea in cluster analysis of the joint application of Ward's and K-means methods. At the same time, it is proposed to modernize each of these methods and supplement them with a third method of splitting/merging pixel clusters. This is useful for cluster analysis of big data described by a convex dependence of the optimal approximation error on the cluster number and also for adjustable object detection in digital image processing, using the optimal hierarchical pixel clustering, which is treated as an alternative to the modern informally defined "semantic" segmentation.

Formation of Fused Images of the Land Surface from Radar and Optical Images in Spatially Distributed On-Board Operational Monitoring Systems.

This paper considers the issues of image fusion in a spatially distributed small-size on-board location system for operational monitoring. The purpose of this research is to develop a new method for the formation of fused images of the land surface based on data obtained from optical and radar devices operated from two-position spatially distributed systems of small aircraft, including unmanned aerial vehicles. The advantages of the method for integrating information from radar and optical information-measuring systems are justified. The combined approach allows removing the limitations of each separate system. The practicality of choosing the integration of information from several widely used variants of heterogeneous sources is shown. An iterative approach is used in the method for combining multi-angle location images. This approach improves the quality of synthesis and increases the accuracy of integration, as well as improves the information content and reliability of the final fused image by using the pixel clustering algorithm, which produces many partitions into clusters. The search for reference points on isolated contours is carried out on a pair of left and right images of the docked image from the selected partition. For these reference points, a functional transformation is determined. Having applied it to the original multi-angle heterogeneous images, the degree of correlation of the fused image is assessed. Both the position of the reference points of the contour and the desired functional transformation itself are refined until the quality assessment of the fusion becomes acceptable. The type of functional transformation is selected based on clustered images and then applied to the original multi-angle heterogeneous images. This process is repeated for clustered images with greater granularity in case if quality assessment of the fusion is considered to be poor. At each iteration, there is a search for pairs of points of the contour of the isolated areas. Areas are isolated with the use of two image segmentation methods. Experiments on the formation of fused images are presented. The result of the research is the proposed method for integrating information obtained from a two-position airborne small-sized radar system and an optical location system. The implemented method can improve the information content, quality, and reliability of the finally established fused image of the land surface.

Intelligent Automatic Segmentation of Wrist Ganglion Cysts Using DBSCAN and Fuzzy C-Means.

Ganglion cysts are common soft tissue masses of the hand and wrist, and small size cysts are often hypoechoic. Thus, identifying them from ultrasonography is not an easy problem. In this paper, we propose an automatic segmentation method using two artificial intelligence algorithms in sequence. A density based unsupervised learning algorithm called DBSCAN is performed as a front-end and its result determines the number of clusters used in the Fuzzy C-Means (FCM) clustering algorithm for quantification of ganglion cyst object. In an experiment using 120 images, the proposed method shows a higher extraction rate (89.2%) and lower false positive rate compared with FCM when the ground truth is set as the human expert's decision. Such human-like behavior is more apparent when the size of the ganglion cyst is small that the quality of ultrasonography is often not very high. With this fully automatic segmentation method, the operator subjectivity that is highly dependent on the experience of the ultrasound examiner can be mitigated with high reliability.

[Left ventricle segmentation in echocardiography based on adaptive mean shift].

The use of echocardiography ventricle segmentation can obtain ventricular volume parameters, and it is helpful to evaluate cardiac function. However, the ultrasound images have the characteristics of high noise and difficulty in segmentation, bringing huge workload to segment the object region manually. Meanwhile, the automatic segmentation technology cannot guarantee the segmentation accuracy. In order to solve this problem, a novel algorithm framework is proposed to segment the ventricle. Firstly, faster region-based convolutional neural network is used to locate the object to get the region of interest. Secondly, K-means is used to pre-segment the image; then a mean shift with adaptive bandwidth of kernel function is proposed to segment the region of interest. Finally, the region growing algorithm is used to get the object region. By this framework, ventricle is obtained automatically without manual localization. Experiments prove that this framework can segment the object accurately, and the algorithm of adaptive mean shift is more stable and accurate than the mean shift with fixed bandwidth on quantitative evaluation. These results show that the method in this paper is helpful for automatic segmentation of left ventricle in echocardiography.

Left ventricle segmentation in echocardiography based on adaptive mean shift / 生物医学工程学杂志

The use of echocardiography ventricle segmentation can obtain ventricular volume parameters, and it is helpful to evaluate cardiac function. However, the ultrasound images have the characteristics of high noise and difficulty in segmentation, bringing huge workload to segment the object region manually. Meanwhile, the automatic segmentation technology cannot guarantee the segmentation accuracy. In order to solve this problem, a novel algorithm framework is proposed to segment the ventricle. Firstly, faster region-based convolutional neural network is used to locate the object to get the region of interest. Secondly, -means is used to pre-segment the image; then a mean shift with adaptive bandwidth of kernel function is proposed to segment the region of interest. Finally, the region growing algorithm is used to get the object region. By this framework, ventricle is obtained automatically without manual localization. Experiments prove that this framework can segment the object accurately, and the algorithm of adaptive mean shift is more stable and accurate than the mean shift with fixed bandwidth on quantitative evaluation. These results show that the method in this paper is helpful for automatic segmentation of left ventricle in echocardiography.

A novel phase-unwrapping method based on pixel clustering and local surface fitting with application to Dixon water-fat MRI.

PURPOSE: To develop and evaluate a novel 2D phase-unwrapping method that works robustly in the presence of severe noise, rapid phase changes, and disconnected regions. THEORY AND METHODS: The MR phase map usually varies rapidly in regions adjacent to wraps. In contrast, the phasors can vary slowly, especially in regions distant from tissue boundaries. Based on this observation, this paper develops a phase-unwrapping method by using a pixel clustering and local surface fitting (CLOSE) approach to exploit different local variation characteristics between the phase and phasor data. The CLOSE approach classifies pixels into easy-to-unwrap blocks and difficult-to-unwrap residual pixels first, and then sequentially performs intrablock, interblock, and residual-pixel phase unwrapping by a region-growing surface-fitting method. The CLOSE method was evaluated on simulation and in vivo water-fat Dixon data, and was compared with phase region expanding labeler for unwrapping discrete estimates (PRELUDE). RESULTS: In the simulation experiment, the mean error ratio by CLOSE was less than 1.50%, even in areas with signal-to-noise ratio equal to 0.5, phase changes larger than π, and disconnected regions. For 350 in vivo knee and ankle images, the water-fat swap ratio of CLOSE was 4.29%, whereas that of PRELUDE was 25.71%. CONCLUSIONS: The CLOSE approach can correctly unwrap phase with high robustness, and benefit MRI applications that require phase unwrapping. Magn Reson Med 79:515-528, 2018. © 2017 International Society for Magnetic Resonance in Medicine.