|
|
Image segmentation using graph neural networks
Boszorád, Matej ; Kolařík, Martin (oponent) ; Myška, Vojtěch (vedoucí práce)
This diploma thesis describes and implements the design of a graph neural network usedfor 2D segmentation of neural structure. The first chapter of the thesis briefly introduces the problem of segmentation. In this chapter, segmentation techniques are divided according to the principles of the methods they use. Each type of technique contains the essence of this category as well as a description of one representative. The second chapter of the diploma thesis explains graph neural networks (GNN for short). Here, the thesis divides graph neural networks in general and describes recurrent graph neural networks(RGNN for short) and graph autoencoders, that can be used for image segmentation, in more detail. The specific image segmentation solution is based on the message passing method in RGNN, which can replace convolution masks in convolutional neural networks.RGNN also provides a simpler multilayer perceptron topology. The second type of graph neural networks characterised in the thesis are graph autoencoders, which use various methods for better encoding of graph vertices into Euclidean space. The last part ofthe diploma thesis deals with the analysis of the problem, the proposal of its specific solution and the evaluation of results. The purpose of the practical part of the work was the implementation of GNN for image data segmentation. The advantage of using neural networks is the ability to solve different types of segmentation by changing training data. RGNN with messaging passing and node2vec were used as implementation GNNf or segmentation problem. RGNN training was performed on graphics cards provided bythe school and Google Colaboratory. Learning RGNN using node2vec was very memory intensive and therefore it was necessary to train on a processor with an operating memory larger than 12GB. As part of the RGNN optimization, learning was tested using various loss functions, changing topology and learning parameters. A tree structure method was developed to use node2vec to improve segmentation, but the results did not confirman improvement for a small number of iterations. The best outcomes of the practical implementation were evaluated by comparing the tested data with the convolutional neural network U-Net. It is possible to state comparable results to the U-Net network, but further testing is needed to compare these neural networks. The result of the thesisis the use of RGNN as a modern solution to the problem of image segmentation and providing a foundation for further research.
|
|
|
Identification and characterization of malicious behavior in behavioral graphs
Varga, Adam ; Burget, Radim (oponent) ; Hajný, Jan (vedoucí práce)
In recent years, there has been an increase in work involving comprehensive malware detection. It is often useful to use a graph format to capture behavior. This is the case with the Avast antivirus program, whose behavioral shield detects malicious behavior and stores it in the form of graphs. Since this is a proprietary solution and Avast antivirus works with its own set of characterized behavior, it was necessary to design our own detection method that will be built on top of these behavioral graphs. This work analyzes graphs of malware behavior captured by the behavioral shield of the Avast antivirus program for the process of deeper detection of malware. Detection of malicious behavior begins with the analysis and abstraction of patterns from the behavioral graph. Isolated patterns can more effectively identify dynamically changing malware. Behavior graphs are stored in the Neo4j graph database and thousands of them are captured every day. The aim of this work was to design an algorithm to identify the behavior of malicious software with emphasis on tagging speed and uniqueness of identified patterns of behavior. Identification of malicious behavior consists in finding the most important properties of trained classifiers and subsequent extraction of a subgraph consisting only of these important properties of nodes and the relationships between them. Subsequently, a rule for the evaluation of the extracted subgraph is proposed. The diploma thesis took place in cooperation with Avast Software s.r.o.
|
|
|
Neural Networks for Automatic Table Recognition
Piwowarski, Lukáš ; Španěl, Michal (oponent) ; Hradiš, Michal (vedoucí práce)
This thesis introduces the reader to the current table recognition techniques mainly used to extract information from historical handwritten and printed tables. We also introduce a method based on graph neural network, which is inspired by the presented techniques. The method consists of three phases: graph initialization, node/edge classification and graph to text transformation phase. In the graph initialization phase, we use the node visibility algorithm and OCR to create an initial graph representation of the input table. In the node and edge classification phase, the nodes and edges are classified, and in the graph to text transformation phase, we fit the graph's nodes into a grid which is then used to produce the final text representation of the table. The implemented model achieved horizontal neighbours detection precision of 68 %, vertical neighbours detection precision of 71 % and cell detection precision of 85 % on the ABP dataset.
|
|
|
Neural Networks for Automatic Table Recognition
Piwowarski, Lukáš ; Španěl, Michal (oponent) ; Hradiš, Michal (vedoucí práce)
This thesis introduces the reader to the current table recognition techniques mainly used to extract information from historical handwritten and printed tables. We also introduce a method based on graph neural network, which is inspired by the presented techniques. The method consists of three phases: graph initialization, node/edge classification and graph to text transformation phase. In the graph initialization phase, we use the node visibility algorithm and OCR to create an initial graph representation of the input table. In the node and edge classification phase, the nodes and edges are classified, and in the graph to text transformation phase, we fit the graph's nodes into a grid which is then used to produce the final text representation of the table. The implemented model achieved horizontal neighbours detection precision of 68 %, vertical neighbours detection precision of 71 % and cell detection precision of 85 % on the ABP dataset.
|
|
|
Identification and characterization of malicious behavior in behavioral graphs
Varga, Adam ; Burget, Radim (oponent) ; Hajný, Jan (vedoucí práce)
In recent years, there has been an increase in work involving comprehensive malware detection. It is often useful to use a graph format to capture behavior. This is the case with the Avast antivirus program, whose behavioral shield detects malicious behavior and stores it in the form of graphs. Since this is a proprietary solution and Avast antivirus works with its own set of characterized behavior, it was necessary to design our own detection method that will be built on top of these behavioral graphs. This work analyzes graphs of malware behavior captured by the behavioral shield of the Avast antivirus program for the process of deeper detection of malware. Detection of malicious behavior begins with the analysis and abstraction of patterns from the behavioral graph. Isolated patterns can more effectively identify dynamically changing malware. Behavior graphs are stored in the Neo4j graph database and thousands of them are captured every day. The aim of this work was to design an algorithm to identify the behavior of malicious software with emphasis on tagging speed and uniqueness of identified patterns of behavior. Identification of malicious behavior consists in finding the most important properties of trained classifiers and subsequent extraction of a subgraph consisting only of these important properties of nodes and the relationships between them. Subsequently, a rule for the evaluation of the extracted subgraph is proposed. The diploma thesis took place in cooperation with Avast Software s.r.o.
|
|
|
Image segmentation using graph neural networks
Boszorád, Matej ; Kolařík, Martin (oponent) ; Myška, Vojtěch (vedoucí práce)
This diploma thesis describes and implements the design of a graph neural network usedfor 2D segmentation of neural structure. The first chapter of the thesis briefly introduces the problem of segmentation. In this chapter, segmentation techniques are divided according to the principles of the methods they use. Each type of technique contains the essence of this category as well as a description of one representative. The second chapter of the diploma thesis explains graph neural networks (GNN for short). Here, the thesis divides graph neural networks in general and describes recurrent graph neural networks(RGNN for short) and graph autoencoders, that can be used for image segmentation, in more detail. The specific image segmentation solution is based on the message passing method in RGNN, which can replace convolution masks in convolutional neural networks.RGNN also provides a simpler multilayer perceptron topology. The second type of graph neural networks characterised in the thesis are graph autoencoders, which use various methods for better encoding of graph vertices into Euclidean space. The last part ofthe diploma thesis deals with the analysis of the problem, the proposal of its specific solution and the evaluation of results. The purpose of the practical part of the work was the implementation of GNN for image data segmentation. The advantage of using neural networks is the ability to solve different types of segmentation by changing training data. RGNN with messaging passing and node2vec were used as implementation GNNf or segmentation problem. RGNN training was performed on graphics cards provided bythe school and Google Colaboratory. Learning RGNN using node2vec was very memory intensive and therefore it was necessary to train on a processor with an operating memory larger than 12GB. As part of the RGNN optimization, learning was tested using various loss functions, changing topology and learning parameters. A tree structure method was developed to use node2vec to improve segmentation, but the results did not confirman improvement for a small number of iterations. The best outcomes of the practical implementation were evaluated by comparing the tested data with the convolutional neural network U-Net. It is possible to state comparable results to the U-Net network, but further testing is needed to compare these neural networks. The result of the thesisis the use of RGNN as a modern solution to the problem of image segmentation and providing a foundation for further research.
|
host ::
RSS.