Evaluating the Readability of Force Directed Graph Layouts: A Deep Learning Approach

Abstract

Existing graph layout algorithms are usually not able to optimize all the aesthetic properties desired in a graph layout. To evaluate how well the desired visual features are reflected in a graph layout, many readability metrics have been proposed in the past decades. However, the calculation of these readability metrics often requires access to the node and edge coordinates and is usually computationally inefficient, especially for dense graphs. Importantly, when the node and edge coordinates are not accessible, it becomes impossible to evaluate the graph layouts quantitatively. In this paper, we present a novel deep learning-based approach to evaluate the readability of graph layouts by directly using graph images. A convolutional neural network architecture is proposed and trained on a benchmark dataset of graph images, which is composed of synthetically-generated graphs and graphs created by sampling from real large networks. Multiple representative readability metrics (including edge crossing, node spread, and group overlap) are considered in the proposed approach. We quantitatively compare our approach to traditional methods and qualitatively evaluate our approach using a case study and visualizing convolutional layers. This work is a first step towards using deep learning based methods to evaluate images from the visualization field quantitatively.

This work has been accepted at IEEE CG&A

References in corpus (6)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Finding community structure in very large networks
• Benchmark graphs for testing community detection algorithms
• What Would a Graph Look Like in This Layout? A Machine Learning Approach to Large Graph Visualization
• Understanding Hidden Memories of Recurrent Neural Networks
• Aesthetic Discrimination of Graph Layouts

Cited by in corpus (15)

• A Survey on ML4VIS: Applying Machine Learning Advances to Data Visualization
• A Deep Generative Model for Graph Layout
• Toward automatic comparison of visualization techniques: Application to graph visualization
• PREVis: Perceived Readability Evaluation for Visualizations
• Quantivine: A Visualization Approach for Large-scale Quantum Circuit Representation and Analysis
• Interactive Visual Pattern Search on Graph Data via Graph Representation Learning
• Rapid GPU-Based Pangenome Graph Layout
• Learning to Automate Chart Layout Configurations Using Crowdsourced Paired Comparison
• Impacts of the Numbers of Colors and Shapes on Outlier Detection: from Automated to User Evaluation
• PSEUDo: Interactive Pattern Search in Multivariate Time Series with Locality-Sensitive Hashing and Relevance Feedback
• Deep Neural Network for DrawiNg Networks, (DNN)^2
• DeepGD: A Deep Learning Framework for Graph Drawing Using GNN
• Evaluating Effects of Background Stories on Graph Perception
• DeepDrawing: A Deep Learning Approach to Graph Drawing
• BigGraphVis: Leveraging Streaming Algorithms and GPU Acceleration for Visualizing Big Graphs