Toward automatic comparison of visualization techniques: Application to graph visualization

Giovannangeli, Loann; Bourqui, Romain; Giot, Romain

doi:10.1016/j.visinf.2020.04.002

Cited by 17 publications

(12 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a survey on quality metrics for information visualization, Behrisch et al [9] stated that DNNs were a promising direction for evaluating representations qualities. Some studies continued exploring this thematic, comparing human subjects to DNNs on the interpretation of simple graphical elements [10] or even complex data structures representations [6], [8]. This field of research is getting more and more interest in the recent years as shown in the Wang et al [11] survey on the application of Machine Learning (ML) techniques to various Information Visualization domains, which they called ML4VIS.…”

Section: A Deep Neural Network For Information Visualizationmentioning

confidence: 99%

“…This field of research is getting more and more interest in the recent years as shown in the Wang et al [11] survey on the application of Machine Learning (ML) techniques to various Information Visualization domains, which they called ML4VIS. Some studies are already defining generic workflows to make use of Deep Learning (DL) models to address some information visualization problems [6], [12]. These methods rely on the assumption that ML techniques could model the perception of graphical content by humans.…”

Section: A Deep Neural Network For Information Visualizationmentioning

confidence: 99%

“…CNNs are better at solving tasks that require to interpret simple elements many times, whereas humans are more capable of interpreting complex representations requiring a higher level of abstraction or prior knowledge. Finally, Giovannangeli et al [6] reproduced two information visualization evaluations of graph representations (i.e., representations of complex data) with CNNs. They were able to draw the same conclusions as the experiments they were reproducing and that were conducted on human subjects.…”

Section: B Deep Neural Network and Humansmentioning

confidence: 99%

“…Therewith, we expect to be able to overcome user evaluations flaws by using DL-based automated approaches. If some works have already shown how DL techniques could be used to help visualization techniques evaluations [6]- [9], only a few measured how Deep Neural Networks (DNNs) are correlated with humans on perception tasks [6], [10].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Analysis of Deep Neural Networks Correlations with Human Subjects on a Perception Task

Giovannangeli

Giot

Benois‐Pineau

et al. 2021

2021 25th International Conference Information Visualisation (IV)

View full text Add to dashboard Cite

Section: A Deep Neural Network For Information Visualizationmentioning

confidence: 99%

Section: A Deep Neural Network For Information Visualizationmentioning

confidence: 99%

Section: B Deep Neural Network and Humansmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Analysis of Deep Neural Networks Correlations with Human Subjects on a Perception Task

Giovannangeli

Giot

Benois‐Pineau

et al. 2021

2021 25th International Conference Information Visualisation (IV)

View full text Add to dashboard Cite

“…1b) provides an overview of the landscapes and object elements (T1). It employs a node-link diagram, which is widely used and studied (Zhao et al 2020;Giovannangeli et al 2020;Shi et al 2020;Pinaud et al 2020), to show not only the elements detected in the current painting, but also all those appear in the data set. We use NetV.js (Han et al 2021) to draw the graph.…”

Section: Element Exploration Modulementioning

confidence: 99%

iPoet: interactive painting poetry creation with visual multimodal analysis

Feng

Chen

Huang

et al. 2021

J Vis

View full text Add to dashboard Cite

Chinese painting poetry is an extraordinary aesthetic phenomenon in world art history. It is not only part of the paintings but also helps us to better understand the spiritual conception that the artists express. In this paper, we present an interactive visual system to enable ordinary users to compose customized painting poetry for ancient Chinese paintings, which contain three properties: (1) We employ object detection and image captioning to describe the scenery depicted in the painting. (2) We extend the modern color theory to analyze the underlying emotions of each painting. (3) We propose an interactive poetry generation method that takes the content description and the emotional expression to add the diversity of the poetry creation. Several visual components are carefully designed to visualize and contextualize the features in the painting. They effectively guide users to steer the creation of personalized painting poems. We conduct efficient case studies and user interviews to demonstrate the effectiveness of our system.

show abstract

Deep Neural Network for DrawiNg Networks, $${(DNN)^{\textit{2}\,}} $$

Giovannangeli

Lalanne

Giot

et al. 2021

Lecture Notes in Computer Science

View full text Add to dashboard Cite

By leveraging recent progress of stochastic gradient descent methods, several works have shown that graphs could be efficiently laid out through the optimization of a tailored objective function. In the meantime, Deep Learning (DL) techniques achieved great performances in many applications. We demonstrate that it is possible to use DL techniques to learn a graph-to-layout sequence of operations thanks to a graph-related objective function. In this paper, we present a novel graph drawing framework called (DNN) 2 : Deep Neural Network for DrawiNg Networks. Our method uses Graph Convolution Networks to learn a model. Learning is achieved by optimizing a graph topology related loss function that evaluates (DNN) 2 generated layouts during training. Once trained, the (DNN) 2 model is able to quickly lay any input graph out. We experiment (DNN) 2 and statistically compare it to optimization-based and regular graph layout algorithms. The results show that (DNN) 2 performs well and are encouraging as the Deep Learning approach to Graph Drawing is novel and many leads for future works are identified.

show abstract

Toward automatic comparison of visualization techniques: Application to graph visualization

Cited by 17 publications

References 16 publications

Analysis of Deep Neural Networks Correlations with Human Subjects on a Perception Task

Analysis of Deep Neural Networks Correlations with Human Subjects on a Perception Task

iPoet: interactive painting poetry creation with visual multimodal analysis

Deep Neural Network for DrawiNg Networks, $${(DNN)^{\textit{2}\,}} $$

Contact Info

Product

Resources

About