maj

[canny.git] / experiments.tex

diff --git a/experiments.tex b/experiments.tex
index aed050525ea5126f9b317d53ae0ca3ef3519d930..0c08d6924a78df9cce07ca0dd0cb947f944d7021 100644 (file)
--- a/experiments.tex
+++ b/experiments.tex
@@ -3,7 +3,7 @@ of the BOSS contest~\cite{Boss10} database is taken.
 In this set, each cover is a $512\times 512$
 grayscale digital image in a RAW format. 
 We restrict experiments to 
 In this set, each cover is a $512\times 512$
 grayscale digital image in a RAW format. 
 We restrict experiments to 

-this set of cover images since this paper is more focussed on 
+this set of cover images since this paper is more focused on 

 the methodology than benchmarking.    
 Our approach is always compared to Hugo~\cite{DBLP:conf/ih/PevnyFB10}
 and to EAISLSBMR~\cite{Luo:2010:EAI:1824719.1824720}.
 the methodology than benchmarking.    
 Our approach is always compared to Hugo~\cite{DBLP:conf/ih/PevnyFB10}
 and to EAISLSBMR~\cite{Luo:2010:EAI:1824719.1824720}.


@@ -69,22 +69,28 @@ The other ones have been designed to tackle this problem.
 \hline
 Schemes & \multicolumn{3}{|c|}{STABYLO} & \multicolumn{2}{|c|}{HUGO}& \multicolumn{2}{|c|}{EAISLSBMR} \\
 \hline
 \hline
 Schemes & \multicolumn{3}{|c|}{STABYLO} & \multicolumn{2}{|c|}{HUGO}& \multicolumn{2}{|c|}{EAISLSBMR} \\
 \hline

-Embedding &   \multicolumn{2}{|c||}{Adaptive} & Fixed & \multicolumn{2}{|c|}{Fixed}& \multicolumn{2}{|c|}{Fixed} \\
+Embedding &   Fixed & \multicolumn{2}{|c|}{Adaptive} &  \multicolumn{2}{|c|}{Fixed}& \multicolumn{2}{|c|}{Fixed} \\

 \hline
 \hline

-Rate &   + STC &  + sample & 10\% & 10\%&6.35\%& 10\%&6.35\%\\ 
+Rate &   10\% &  + sample & + STC  &  10\%&6.35\%& 10\%&6.35\%\\ 

 \hline
 \hline

-PSNR &  66.55 (\textbf{-0.8\%}) & 63.48  & 61.86  & 64.65 & {67.08} & 60.8 & 62.9\\ 
+PSNR & 61.86 & 63.48 &  66.55 (\textbf{-0.8\%})     & 64.65 & {67.08} & 60.8 & 62.9\\ 

 \hline
 \hline

-PSNR-HVS-M & 78.6 (\textbf{-0.8\%})  & 75.39 & 72.9 & 76.67 & {79.23} & 61.3  & 63.4\\ 
+PSNR-HVS-M & 72.9 & 75.39 & 78.6 (\textbf{-0.8\%})    & 76.67 & {79.23} & 61.3  & 63.4\\ 

 %\hline
 %BIQI & 28.3 & 28.28 & 28.4 & 28.28 & 28.28 & 28.2 & 28.2\\ 
 \hline
 %\hline
 %BIQI & 28.3 & 28.28 & 28.4 & 28.28 & 28.28 & 28.2 & 28.2\\ 
 \hline

-wPSNR & 86.43(\textbf{-1.6\%}) & 80.59 & 77.47& 83.03 & {87.8} & & 80.6\\ 
+wPSNR & 77.47 & 80.59 & 86.43(\textbf{-1.6\%})  & 83.03 & {87.8} & 76.7 & 80.6\\ 

 \hline
 \end{tabular}
 \hline
 \end{tabular}

+
+\begin{footnotesize}
+\vspace{2em}
+Variances given in bold font express the quality differences between 
+HUGO and STABYLO with  STC+adaptive parameters.
+\end{footnotesize}
+

 \end{center}
 \caption{Quality Measures of Steganography Approaches\label{table:quality}}
 \end{center}
 \caption{Quality Measures of Steganography Approaches\label{table:quality}}

-\label{table:quality}

 \end{table*}
 
 
 \end{table*}
 
 


@@ -101,7 +107,7 @@ into the edge detection.
 Let us focus on the quality of HUGO images: with a given fixed 
 embedding rate (10\%), 
 HUGO always produces images whose quality is higher than the STABYLO's one.
 Let us focus on the quality of HUGO images: with a given fixed 
 embedding rate (10\%), 
 HUGO always produces images whose quality is higher than the STABYLO's one.

-However our appraoch always outperforms EAISLSBMR since this one may modify 
+However our approach always outperforms EAISLSBMR since this one may modify 

 the two least significant bits whereas STABYLO only alter LSB.
 
 If we combine \emph{adaptive} and \emph{STC} strategies 
 the two least significant bits whereas STABYLO only alter LSB.
 
 If we combine \emph{adaptive} and \emph{STC} strategies 


@@ -126,7 +132,7 @@ give quality metrics for fixed embedding rates from a large base of images.
 
 
 
 
 
 

-The quality of our approach has been evaluated through the two 
+The steganalysis quality of our approach has been evaluated through the two 

 AUMP~\cite{Fillatre:2012:ASL:2333143.2333587}
 and Ensemble Classifier~\cite{DBLP:journals/tifs/KodovskyFH12} based steganalysers.
 Both aims at detecting hidden bits in grayscale natural images and are 
 AUMP~\cite{Fillatre:2012:ASL:2333143.2333587}
 and Ensemble Classifier~\cite{DBLP:journals/tifs/KodovskyFH12} based steganalysers.
 Both aims at detecting hidden bits in grayscale natural images and are 


@@ -141,7 +147,6 @@ machine learning step, which is often
 implemented as support vector machine,
 can be favorably executed thanks to an ensemble classifier.
 
 implemented as support vector machine,
 can be favorably executed thanks to an ensemble classifier.
 

-%citer le second tableau, comparer avec EAISLSBMR

 
 \begin{table*}
 \begin{center}
 
 \begin{table*}
 \begin{center}


@@ -150,13 +155,13 @@ can be favorably executed thanks to an ensemble classifier.
 \hline
 Schemes & \multicolumn{3}{|c|}{STABYLO} & \multicolumn{2}{|c|}{HUGO}& \multicolumn{2}{|c|}{EAISLSBMR}\\
 \hline
 \hline
 Schemes & \multicolumn{3}{|c|}{STABYLO} & \multicolumn{2}{|c|}{HUGO}& \multicolumn{2}{|c|}{EAISLSBMR}\\
 \hline

-Embedding &   \multicolumn{2}{|c|}{Adaptive} & Fixed & \multicolumn{2}{|c|}{Fixed}& \multicolumn{2}{|c|}{Fixed} \\
+Embedding & Fixed &   \multicolumn{2}{|c|}{Adaptive}  & \multicolumn{2}{|c|}{Fixed}& \multicolumn{2}{|c|}{Fixed} \\

 \hline
 \hline

-Rate &   + STC &  + sample & 10\% & 10\%& 6.35\%& 10\%& 6.35\%\\ 
+Rate & 10\% &  + sample &   + STC   & 10\%& 6.35\%& 10\%& 6.35\%\\ 

 \hline
 \hline

-AUMP & 0.39  & 0.33  & 0.22     &  0.50 &  0.50 & 0.49 & 0.50 \\
+AUMP & 0.22 & 0.33 & 0.39         &  0.50 &  0.50 & 0.49 & 0.50 \\

 \hline
 \hline

-Ensemble Classifier & \textbf{0.47} & 0.44 & 0.35     & 0.48 &  0.49  &  0.43  & 0.46 \\
+Ensemble Classifier & 0.35 & 0.44 & 0.47       & 0.48 &  0.49  &  0.43  & 0.46 \\

 
 \hline
 \end{tabular}
 
 \hline
 \end{tabular}


@@ -166,8 +171,14 @@ Ensemble Classifier & \textbf{0.47} & 0.44 & 0.35     & 0.48 &  0.49  &  0.43  &
 \end{table*}
 
 
 \end{table*}
 
 

-Results show that our approach is more easily detectable than HUGO, which
-is the most secure steganographic tool, as far as we know. However due to its 
+Results are summarized in Table~\ref{table:steganalyse}.
+First of all, STC outperforms the sample strategy for the two steganalysers, as 
+already noticed in the quality analysis presented in the previous section. 
+Next, our approach is more easily detectable than HUGO, which
+is the most secure steganographic tool, as far as we know. 
+However by combining \emph{adaptive} and \emph{STC} strategies
+our approach obtains similar results than HUGO ones.
+However due to its 

 huge number of features integration, it is not lightweight, which justifies 
 in the authors' opinion the consideration of the proposed method.   
 
 huge number of features integration, it is not lightweight, which justifies 
 in the authors' opinion the consideration of the proposed method.