erwan/phd-thesis
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Better figures

Browse Source
This commit is contained in:
Erwan Normand
2024-09-24 15:07:13 +02:00
parent 9ba5d344a5
commit b8b799df3d
18 changed files with 383 additions and 418 deletions
Download Patch File Download Diff File Expand all files Collapse all files

91
2-perception/ar-textures/3-results.tex
View File

@@ -7,17 +7,16 @@
\subsubsection{Confusion Matrix}
\label{results_matching_confusion_matrix}
\begin{subfigs}{results_matching_ranking}{%
(Left) Confusion matrix of the matching task, with the presented visual textures as columns and the selected haptic texture in proportion as rows. %
The number in a cell is the proportion of times the corresponding haptic texture was selected in response to the presentation of the corresponding visual texture. %
The diagonal represents the expected correct answers. %
Holm-Bonferroni adjusted binomial test results are marked in bold when the proportion is higher than chance (\ie more than 11~\%, \pinf{0.05}).
%
(Right) Means with bootstrap 95~\% confidence interval of the three rankings of the haptic textures alone, the visual textures alone, and the visuo-haptic texture pairs. %
A lower rank means that the texture was considered rougher, a higher rank means smoother. %
}
\subfig[0.58]{results/matching_confusion_matrix}%
\subfig[0.41]{results/ranking_mean_ci}%
\begin{subfigs}{results_matching_ranking}{Results of the matching and ranking tasks. }[][
\item Confusion matrix of the matching task, with the presented visual textures as columns and the selected haptic texture in proportion as rows.
The number in a cell is the proportion of times the corresponding haptic texture was selected in response to the presentation of the corresponding visual texture.
The diagonal represents the expected correct answers.
Holm-Bonferroni adjusted binomial test results are marked in bold when the proportion is higher than chance (\ie more than 11~\%, \pinf{0.05}).
\item Means with bootstrap 95~\% confidence interval of the three rankings of the haptic textures alone, the visual textures alone, and the visuo-haptic texture pairs.
A lower rank means that the texture was considered rougher, a higher rank means smoother.
]
\subfig[0.58]{results/matching_confusion_matrix}%
\subfig[0.41]{results/ranking_mean_ci}%
\end{subfigs}
\figref{results_matching_ranking} (left) shows the confusion matrix of the matching task with the visual textures and the proportion of haptic texture selected in response, \ie the proportion of times the corresponding haptic texture was selected in response to the presentation of the corresponding visual texture.
@@ -57,7 +56,6 @@ Normality was verified with a QQ-plot of the model residuals.
%
No statistical significant effect of \textit{Visual Texture} was found (\anova{8}{512}{1.9}, \p{0.06}) on \textit{Completion Time} (\geomean{44}{\s}, \ci{42}{46}), indicating an equal difficulty and participant behaviour for all the visual textures.
\subsection{Textures Ranking}
\label{results_ranking}
@@ -81,28 +79,27 @@ A Wilcoxon signed-rank test indicated that this difference was statistically sig
%
These results indicate, with \figref{results_matching_ranking} (right), that the two haptic and visual modalities were integrated together, the resulting roughness ranking being between the two rankings of the modalities alone, but with haptics predominating.
\subsection{Perceived Similarity of Visual and Haptic Textures}
\label{results_similarity}
\begin{subfigs}{results_similarity}{%
(Left) Correspondence analysis of the matching task confusion matrix (\figref{results_matching_ranking}, left).
The visual textures are represented as blue squares, the haptic textures as red circles. %
The closer the textures are, the more similar they were judged. %
The first dimension (horizontal axis) explains 60~\% of the variance, the second dimension (vertical axis) explains 30~\% of the variance.
(Right) Dendrograms of the hierarchical clusterings of the haptic textures (left) and visual textures (right) of the matching task confusion matrix (\figref{results_matching_ranking}, left), using Euclidian distance and Ward's method. %
The height of the dendrograms represents the distance between the clusters. %
}
\begin{minipage}[c]{0.50\linewidth}%
\centering%
\subfig[1.0]{results/matching_correspondence_analysis}%
\end{minipage}%
\begin{minipage}[c]{0.50\linewidth}%
\centering%
\subfig[0.66]{results/clusters_haptic}%
\par%
\subfig[0.66]{results/clusters_visual}%
\end{minipage}%
(Left) Correspondence analysis of the matching task confusion matrix (\figref{results_matching_ranking}, left).
The visual textures are represented as blue squares, the haptic textures as red circles. %
The closer the textures are, the more similar they were judged. %
The first dimension (horizontal axis) explains 60~\% of the variance, the second dimension (vertical axis) explains 30~\% of the variance.
(Right) Dendrograms of the hierarchical clusterings of the haptic textures (left) and visual textures (right) of the matching task confusion matrix (\figref{results_matching_ranking}, left), using Euclidian distance and Ward's method. %
The height of the dendrograms represents the distance between the clusters. %
}
\begin{minipage}[c]{0.50\linewidth}%
\centering%
\subfig[1.0]{results/matching_correspondence_analysis}%
\end{minipage}%
\begin{minipage}[c]{0.50\linewidth}%
\centering%
\subfig[0.66]{results/clusters_haptic}%
\par%
\subfig[0.66]{results/clusters_visual}%
\end{minipage}%
\end{subfigs}
The high level of agreement between participants on the three haptic, visual and visuo-haptic rankings (\secref{results_ranking}), as well as the similarity of the within-participant rankings, suggests that participants perceived the roughness of the textures similarly, but differed in their strategies for matching the haptic and visual textures in the matching task (\secref{results_matching}).
@@ -131,12 +128,15 @@ The five identified visual texture clusters were: "Roughest" \{Metal Mesh\}; "Ro
%
They are also easily identifiable on the visual ranking results, which also made it possible to name them.
\begin{subfigs}{results_clusters}{%
(Left) Confusion matrix of the visual texture clusters with the corresponding haptic texture clusters selected in proportion. %
(Right) Confusion matrix of the visual texture ranks with the corresponding haptic texture clusters selected in proportion. %
(Both) Holm-Bonferroni adjusted binomial test results are marked in bold when the proportion is higher than chance (\ie more than 20~\%, \pinf{0.05}).
}
\subfig[1]{results/haptic_visual_clusters_confusion_matrices}%
\begin{subfigs}{results_clusters}{
Confusion matrices of the visual textures with the corresponding haptic texture clusters selected in proportion.
}[
Holm-Bonferroni adjusted binomial test results are marked in bold when the proportion is higher than chance (\ie more than 20~\%, \pinf{0.05}).
][
\item Confusion matrix of the visual texture clusters.
\item Confusion matrix of the visual texture ranks.
]
\subfig[1]{results/haptic_visual_clusters_confusion_matrices}%
\end{subfigs}
Based on these results, two alternative confusion matrices were constructed.
@@ -153,19 +153,18 @@ A two-sample Pearson Chi-Squared test (\chisqr{24}{540}{342}, \pinf{0.001}) and
%
This shows that the participants consistently identified the roughness of each visual texture and selected the corresponding haptic texture cluster.
\subsection{Questionnaire}
\label{results_questions}
\begin{subfigs}{results_questions}{%
Boxplots of the 7-item Likert scale question results (1=Not at all, 7=Extremely) %
with Holm-Bonferroni adjusted pairwise Wilcoxon signed-rank tests %
(*** is \pinf{0.001} and ** is \pinf{0.01}),
by modality (left) and by task (right). %
Lower is better for Difficulty and Uncomfortable; higher is better for Realism and Textures Match.%
}
\subfig[0.32]{results/questions_modalities}%
\subfig[0.49]{results/questions_tasks}%
\begin{subfigs}{results_questions}{Boxplots of the questionnaire results for each visual hand rendering.}[
Pairwise Wilcoxon signed-rank tests with Holm-Bonferroni adjustment: * is \pinf{0.05}, ** is \pinf{0.01} and *** is \pinf{0.001}.
Lower is better for Difficulty and Uncomfortable; higher is better for Realism and Textures Match.%
][
\item By modality.
\item By task.
]
\subfig[0.32]{results/questions_modalities}%
\subfig[0.49]{results/questions_tasks}%
\end{subfigs}
\figref{results_questions} presents the questionnaire results of the matching and ranking tasks.