Attributed Relational Graph Research Articles

Motif-Based Graph Representation Learning with Application to Chemical Molecules

This work considers the task of representation learning on the attributed relational graph (ARG). Both the nodes and edges in an ARG are associated with attributes/features allowing ARGs to encode rich structural information widely observed in real applications. Existing graph neural networks offer limited ability to capture complex interactions within local structural contexts, which hinders them from taking advantage of the expression power of ARGs. We propose motif convolution module (MCM), a new motif-based graph representation learning technique to better utilize local structural information. The ability to handle continuous edge and node features is one of MCM’s advantages over existing motif-based models. MCM builds a motif vocabulary in an unsupervised way and deploys a novel motif convolution operation to extract the local structural context of individual nodes, which is then used to learn higher level node representations via multilayer perceptron and/or message passing in graph neural networks. When compared with other graph learning approaches to classifying synthetic graphs, our approach is substantially better at capturing structural context. We also demonstrate the performance and explainability advantages of our approach by applying it to several molecular benchmarks.

Automatic detection of fault-controlled rivers using spatial pattern matching

The river system of a region records its tectonic evolution. This study expects to infer some potential faults by effectively quantifying river morphology and identifying rivers with specific morphological and structural characteristics. Thus, this study aims to develop a novel method for automatically detecting fault-controlled rivers (FCRs) from drainage maps using spatial pattern matching and to support the scientific inference of potential faults. The method involves (1) constructing a scene model for the entire drainage basin in the study area based on the attributed relational graph (ARG) model; (2) defining spatial patterns for four types of FCRs using the ARG model, including straight river reaches, right-angle river reaches, barb rivers, and contra-aperture rivers, and storing all river patterns in a spatial pattern library; and (3) detecting the four types of FCRs with a spatial pattern-matching algorithm and the FCR pattern library. Case studies demonstrated the basic effectiveness of this method for detecting FCRs in the study areas of Bomi County (Parong Tsangpo River Valley and Yigong Zangbo River Valley) and southeast of the Qinghai Tibet Plateau. The goals of this study are to quantify river shape better, support the preliminary detection of potential faults, and provide a useful reference for structure-based spatial queries in GIS. Additionally, more river pattern types with specific morphological and structural characteristics can be automatically detected through dynamic maintenance of the spatial pattern library.

Automatic detection of geological folds using attributed relational graphs and formal grammar

The accurate detection and characterization of geological structures are of primary importance for tectonic trend analysis, mineral and water resource exploration, major engineering projects, and geological disaster prevention. However, interpretation and mapping of geological structures by means of manual interpretation are inefficient and high-cost, and the performance varies greatly with the experts’ ability and experience. Folds are complex geological structures, and one of the most common and impressive geological phenomena. The objective of this study is to develop a novel method for the automatic detection of folds from geological maps using attributed relational graphs (ARG) and formal grammar. The method mainly involves (1) partitioning a study area into detecting units by the potential core stratum of fold (CSF) using Voronoi generator algorithm, (2) generating a strip for each detected unit used as modeling domain, (3) using ARG modeling to describe the spatial relations of strata for each strip, (4) defining the patterns of repeated strata corresponding to folds using formal grammar, (5) generating strings from the ARG model, and recognizing whether a symmetrically repeated substring can be extracted from the string using grammar inference to detect the presence of folds, (6) identifying fold types according to the temporal relationships of strata, and (7) mapping folds. Case studies demonstrate the effectiveness of this method for detecting folds in the Mount Lu and Muyunling areas. The proposed approach allows to speed up the interpretation of synclines/anticlines in providing a first rough detection solution, and can provide an appropriate solution to facilitate the detection of three-dimensional geological features from two-dimensional data.

Visual graph mining for graph matching

In this study, we formulate the concept of “mining maximal-size frequent subgraphs” in the challenging domain of visual data (images and videos). In general, visual knowledge can usually be modeled as attributed relational graphs (ARGs) with local attributes representing local parts and pairwise attributes describing the spatial relationship between parts. Thus, from a practical perspective, such mining of maximal-size subgraphs can be regarded as the discovery of common objects from visual data without given annotations of object bounding boxes. From a theoretical perspective, in this study, we propose a generic definition of common subgraphs among ARGs. Many previous studies can be roughly considered as special cases of the definition. In our definition, we consider 1) variations of unary/pairwise attributes among different ARGs, 2) linkage conditions of different nodes, and 3) the learning of similarity metrics for each node. The generality of our subgraph pattern proposes great challenges to the graph-mining algorithm. We propose an approximate but efficient solution to the mining problem. We conduct five experiments to evaluate our method with different kinds of visual data, including videos and RGB/RGB-D images. These experiments demonstrate the generality of the proposed method.

Typification for Façade Structures Based on User Perception

Typification is a well-established operator of map generalization. Although it is widely used in many existing research fields, less discussion has been devoted to the quality of typification. This paper presents a user survey for the evaluation of different typification results of façade structures under different constraints. The survey shows that preservation of the shape of the features is the most important constraint for a reasonable typification process, which has also been quantitatively verified by calculating the similarities between the typified façades and the original façade using attributed relational graph (ARG) and nested earth mover’s distance (NEMD) algorithms. Based on that, an algorithm is developed to generate perceivably reasonable representation from the original facade with decreasing map scale. The algorithm is implemented and tested on a number of façades. Experiments reveal that the typification can be automatically conducted and can create results which are well associated with the original façades.

Exploring structure for long-term tracking of multiple objects in sports videos

In this paper, we propose a novel approach for exploiting structural relations to track multiple objects that may undergo long-term occlusion and abrupt motion. We use a model-free approach that relies only on annotations given in the first frame of the video to track all the objects online, i.e. without knowledge from future frames. We initialize a probabilistic Attributed Relational Graph (ARG) from the first frame, which is incrementally updated along the video. Instead of using the structural information only to evaluate the scene, the proposed approach considers it to generate new tracking hypotheses. In this way, our method is capable of generating relevant object candidates that are used to improve or recover the track of lost objects. The proposed method is evaluated on several videos of table tennis, volleyball, and on the ACASVA dataset. The results show that our approach is very robust, flexible and able to outperform other state-of-the-art methods in sports videos that present structural patterns.

A Multi-Scale Settlement Matching Algorithm Based on ARG

Homonymous entity matching is an important part of multi-source spatial data integration, automatic updating and change detection. Considering the low accuracy of existing matching methods in dealing with matching multi-scale settlement data, an algorithm based on Attributed Relational Graph (ARG) is proposed. The algorithm firstly divides two settlement scenes at different scales into blocks by small-scale road network and constructs local ARGs in each block. Then, ascertains candidate sets by merging procedures and obtains the optimal matching pairs by comparing the similarity of ARGs iteratively. Finally, the corresponding relations between settlements at large and small scales are identified. At the end of this article, a demonstration is presented and the results indicate that the proposed algorithm is capable of handling sophisticated cases.

A Multi-Scale Settlement Matching Algorithm Based on ARG

Abstract. Homonymous entity matching is an important part of multi-source spatial data integration, automatic updating and change detection. Considering the low accuracy of existing matching methods in dealing with matching multi-scale settlement data, an algorithm based on Attributed Relational Graph (ARG) is proposed. The algorithm firstly divides two settlement scenes at different scales into blocks by small-scale road network and constructs local ARGs in each block. Then, ascertains candidate sets by merging procedures and obtains the optimal matching pairs by comparing the similarity of ARGs iteratively. Finally, the corresponding relations between settlements at large and small scales are identified. At the end of this article, a demonstration is presented and the results indicate that the proposed algorithm is capable of handling sophisticated cases.

Object Discovery: Soft Attributed Graph Mining.

We categorize this research in terms of its contribution to both graph theory and computer vision. From the theoretical perspective, this study can be considered as the first attempt to formulate the idea of mining maximal frequent subgraphs in the challenging domain of messy visual data, and as a conceptual extension to the unsupervised learning of graph matching. We define a soft attributed pattern (SAP) to represent the common subgraph pattern among a set of attributed relational graphs (ARGs), considering both their structure and attributes. Regarding the differences between ARGs with fuzzy attributes and conventional labeled graphs, we propose a new mining strategy that directly extracts the SAP with the maximal graph size without applying node enumeration. Given an initial graph template and a number of ARGs, we develop an unsupervised method to modify the graph template into the maximal-size SAP. From a practical perspective, this research develops a general platform for learning the category model (i.e., the SAP) from cluttered visual data (i.e., the ARGs) without labeling "what is where," thereby opening the possibility for a series of applications in the era of big visual data. Experiments demonstrate the superior performance of the proposed method on RGB/RGB-D images and videos.

Semantic Indexing based Remote Sensing Image Retrieval: An Intelligent Decomposition Approach

formulating physical contact with the object, Remote sensing is the achievement of information about an object or phenomenon. The contact of remote sensing (RS) images will turn into more complicated due to the vast data quantity, to defeat this challenges the users can admittance remote sensing images based on semantics. In the existing method, the Quin- tree is used for the decomposition of Content Based Image Retrieval in Remote Sensing, but it has poor retrieval accuracy. So the intelligent decomposition phase is used in our proposed method, which decomposes the image based on the spatial- spectral heterogeneity. The proposed method will perform visual feature, object semantic, spatial relationship semantic, scene semantic based retrievals to ensure fine retrieval schema, which will obtained by applying mapping and the SS modelling in the decomposed remote sensing image. The human intervention will be introduced in the system to ensure the high retrieval accuracy. The implementation result shows the effectiveness of proposed technique, in segmenting the text lines from the input document. The performance of the proposed method is evaluated by comparing the result of proposed method with the conventional SBRSIR technique. The comparison result shows that our proposed method more accurately retrieves the images based on the VF, OS and SS than the conventional SBRSIR technique. KeywordsRemote sensing (RS) images, Visual Features (VF), Object Semantics, Watershed Segmentation, SVM, Attribute Relational Graph (ARG), Spatial-spectral Heterogeneity, Scene Matching model 1. INTRODUCTIONan escalating rate, Image database systems are being

Tracking Multiple Persons under Occlusion Using Appearance Model Based on Attributed Relational Graph

Appearance features are important for tracking persons in stationary scenes. The proposed algorithm was based on appearance mold built by attributed relational graph (ARG). The ARG was used for modeling human appearance features containing color and spatial information. The matching degree of the ARGs was utilized for analyzing tracking situations in current frame. For tracking persons under occlusion, multiple feature patches were generated and the genetic algorithm was used for finding optimal labels of patches. Experiments showed the utility and performance of the proposed approach.

A New Method of Medical Image Retrieval for Computer-Aided Diagnosis

In the field of computer-aided diagnosis the topics of image retrieval is an important approach. According to the difference of retrieval technology, modeling spatial context (e.g., autocorrelation) is a key challenge in image classification and retrieval problems that arise in image regions. This work proposes a new approach to the retrieval of medical images from traditional Markov Random Field model. Contrasting with previous work, this method relies on coping with the ambiguity of spatial relative position concepts: a new definition of the geometric relationship between two objects in a fuzzy set framework is proposed. Furthermore, Fuzzy Attributed Relational Graphs (FARGs) are used in this framework, where each node represents an image object and each edge represents the relationship between two objects. The generalization performance of this approach is then compared with alternative models over the IRMA dataset. These experiments show that our method outperforms the traditional models, such as MRF, FGM, SVM e.g., in terms of several standard measures.

An efficient clustering-based retrieval framework for real crime scene footwear marks

As the most common type of evidence at crime scenes, footwear marks are found more often than fingerprints, and yet left largely unused due to lack of efficient and reliable tools. While the central task is stated simply – retrieve the closest matches among a database of known outsole prints – the difficulty is the poor quality of the marks and a very large and increasing number of outsole patterns. Since grouping the database into clusters can dramatically speed-up retrieval, we propose clustering based on recurring outsole patterns. The clustered database is used to retrieve similar prints for a given crime scene mark. Geometric shapes like line segments, circles and ellipses are proposed as features for crime scene marks. Then these features are structurally represented in the form of an attributed relational graph (ARG). Robust ARG matching is achieved with the introduced footwear print distance (FPD), a similarity measure for footwear prints. Sensitivity analysis of FPD is performed to show its robustness. The proposed system is invariant to scale, translation, rotation and insensitive to noise and degradations of the prints. Experiments show that the approach outperforms other state-of-the-art footwear print retrieval systems.

Detection and typification of linear structures for dynamic visualization of 3D city models

Cluttering is a fundamental problem in 3D city model visualization. In this paper, a novel method for removing cluttering by typification of linear building groups is proposed. This method works in static as well as dynamic visualization of 3D city models. The method starts by converting building models in higher Levels of Details (LoDs) into LoD1 with ground plan and height. Then the Minimum Spanning Tree (MST) is generated according to the distance between the building ground plans. Based on the MST, linear building groups are detected for typification. The typification level of a building group is determined by its distance to the viewpoint as well as its viewing angle. Next, the selected buildings are removed and the remaining ones are adjusted in each group separately. To preserve the building features and their spatial distribution, Attributed Relational Graph (ARG) and Nested Earth Mover’s Distance (NEMD) are used to evaluate the difference between the original building objects and the generalized ones. The experimental results indicate that our method can reduce the number of buildings while preserving the visual similarity of the urban areas.

Partial retrieval of CAD models based on the gradient flows in Lie group

Based on the gradient flows in Lie group, a partial retrieval approach for CAD models is presented in this paper. First, a representation of the face Attributed Relational Graph (ARG) for a CAD model is created from its B-rep model and thus partial retrieval is converted to a subgraph matching problem. Then, an optimization method is adopted to solve the matching problem, where the optimization variable is the vertex mapping and the objective function is the measurement of compatibility between the mapped vertices and between the mapped edges. Different from most previously proposed methods, a homogeneous transformation matrix is introduced to represent the vertex mapping in subgraph matching, whose translational sub-matrix gives the vertex selection in the larger graph and whose orthogonal sub-matrix presents the vertex permutation for the same-sized mapping from the selected vertices to the smaller graph's vertices. Finally, a gradient flow method is developed to search for optimal matching matrix in Special Euclidean group SE ( n ). Here, a penalty approach is used to handle the constraints on the elements of the matching matrix, which leads its orthogonal part to be a permutation matrix and its translational part to have different integer elements. Experimental results show that it is a promising method to support the partial retrieval of CAD models. ► Partial retrieval of CAD models is converted to a subgraph matching problem. ► The match is represented by a homogeneous transformation matrix. ► The translation part of the matrix gives the vertex selection in the larger graph. ► Its rotation part expresses the permutation of the selected vertices in a match. ► A gradient flow method is used to search for the optimal matching matrix in SE ( n ).

TRACKING MULTIPLE PERSONS BASED ON ATTRIBUTED RELATIONAL GRAPH

The appearance model is very effective in tracking multiple persons. The main difficulty in tracking persons is to represent appearance reliably and effectively, especially in the presence of occlusions. In this paper, an effective Attributed Relational Graph (ARG) based tracking algorithm is presented to track multiple persons even under occlusions. The appearance of each person is expressed by an ARG model which not only combines color feature with spatial information but also illustrates the relations among body parts. The similarity of ARG models is computed to build a matching matrix in consecutive frames. Four tracking situations are determined according to the matching matrix. In addition, to track persons under occlusions, probabilistic relaxation labeling in the ARG models of body parts is deduced to label occluded persons optimally. Experimental validation of the proposed tracking method is verified and presented on indoor and outdoor sequences.

Medical Image Retrieval by Fuzzy Set Based Geometric Relationships

Modeling spatial context (e.g., autocorrelation) is a key challenge in classification and retrieval problems that arise in image processing regions. This work proposes a new approach for medical images retrieval enlightened by traditional Markov Random Field model and improve on it. Contrasting with previous work, this method relies on coping with the ambiguity of spatial relative position concepts: a new definition of the geometric relationship between two objects in a fuzzy set framework is proposed. This definition is based on a fuzzy pattern-matching approach, which comparing an object by the fuzzy set representation of the degree of position satisfaction to a reference object. Furthermore, Fuzzy Attributed Relational Graphs (FARGs) are used in this framework for the purpose of medical image similarity measurement.

Automatic Online Educational Game Content Creation by Identifying Similar Chinese Characters with Radical Extraction and Graph Matching Algorithms

One of the difficulties in learning Chinese characters is distinguishing similar characters. This can cause misunderstanding and miscommunication in daily life. Thus, it is important for students learning the Chinese language to be able to distinguish similar characters and understand their proper usage. In this paper, the authors propose a game style framework to train students to distinguish similar characters. A major component in this framework is the search for similar Chinese characters in the system. From the authors’ prior work, they find the similar characters by the radical information and stroke correspondence determination. This paper improves the stroke correspondence determination by using the attributed relational graph (ARG) matching algorithm that considers both the stroke and spatial relationship during matching. The experimental results show that the new proposed method is more accurate in finding similar Chinese characters. Additionally, the authors have implemented online educational games to train students to distinguish similar Chinese characters and made use of the improved matching method for creating the game content automatically.

Practical algorithms for subgroup detection in covert networks

In this paper, we present algorithms for subgroup detection and demonstrated them with a real-time case study of USS Cole bombing terrorist network. The algorithms are demonstrated in an application by a prototype system. The system finds associations between terrorist and terrorist organisations and is capable of determining links between terrorism plots occurred in the past, their affiliation with terrorist camps, travel record, funds transfer, etc. The findings are represented by a network in the form of an Attributed Relational Graph (ARG). Paths from a node to any other node in the network indicate the relationships between individuals and organisations. The system also provides assistance to law enforcement agencies, indicating when the capture of a specific terrorist will more likely destabilise the terrorist network. In this paper, we discuss the important application area related to subgroups in a terrorist cell using filtering of graph algorithms. The novelty of the algorithms can be easily found from the results they produce.

Attributed relational graph matching based on the nested assignment structure

In this paper, we propose a new ARG matching scheme based on the nested assignment structure to assess the similarity between two attributed relational graphs (ARGs). ARGs are represented by nodes and edges containing unary attributes and binary relations between nodes, respectively. The nested assignment structure consists of inner and outer steps. In the inner step, to form a distance matrix, combinatorial differences between every pair of nodes in two ARGs are computed by using an assignment algorithm. Then, in the outer step, a correspondence between nodes in the two ARGs is established by using an assignment algorithm based on the distance matrix. The proposed ARG matching scheme consists of three procedures as follows: first, in the initializing procedure, the nested assignment structure is performed to generate an initial correspondence between nodes in two ARGs. Next, the correspondence is refined by iteratively performing the updating procedure, which also utilizes the nested assignment structure, until the correspondence does not change. Finally, the verifying procedure can be performed in case that some nodes to be matched in the two ARGs are missing. From experimental results, the proposed ARG matching scheme shows superior matching performance and localizes target objects robustly and correctly even in severely noisy and occluded scenes.