no code implementations • 29 Jul 2024 • Hyeon Yu, Jenny Benois-Pineau, Romain Bourqui, Romain Giot, Alexey Zhukov
This paper investigates the use of Mean Opinion Score (MOS), a common image quality metric, as a user-centric evaluation metric for XAI post-hoc explainers.
1 code implementation • CBMI 2023 2023 • Luc-Etienne Pommé, Romain Bourqui, Romain Giot
We also propose two new pairs of metrics that overcome some evaluation issues: (a) Insertion and Deletion Spearman correlation coefficients which both estimate a correlation between the computed scores in a saliency map and the importance for the model of the associated pixels in the image.
no code implementations • Computer Graphics Forum 2023 • Biagio La Rosa, Graziano Blasilli, Romain Bourqui, David Auber, Giuseppe Santucci, Roberto Capobianco, Enrico Bertini, Romain Giot, Marco Angelini
The survey concludes by identifying future research challenges and bridging activities that are helpful to strengthen the role of Visual Analytics as effective support for eXplainable Deep Learning and to foster the adoption of Visual Analytics solutions in the eXplainable Deep Learning community.
1 code implementation • 19 Aug 2022 • Loann Giovannangeli, Frederic Lalanne, Romain Giot, Romain Bourqui
While many graph drawing algorithms consider nodes as points, graph visualization tools often represent them as shapes.
1 code implementation • 8 Aug 2021 • Loann Giovannangeli, Frederic Lalanne, David Auber, Romain Giot, Romain Bourqui
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.
no code implementations • 10 Mar 2021 • Loann Giovannangeli, Romain Giot, David Auber, Romain Bourqui
When encoded with one attribute, the difficulty depends on that attribute heterogeneity until its capacity limit (7 for color, 5 for shape) is reached.
2 code implementations • Tenth International Conference on Image Processing Theory, Tools and Applications 2020 • Kazi Ahmed Asif Fuad, Pierre-Etienne Martin (1, 2), Romain Giot, Romain Bourqui, Jenny Benois-Pineau, Akka Zemmari
Features visualization is performed at the RGB and Optical flow branches of the network.