Mercurial > ift6266
comparison writeup/nips2010_cameraready.tex @ 606:bd7d11089a47
more fix to newpage command.
| author | Frederic Bastien <nouiz@nouiz.org> |
|---|---|
| date | Mon, 22 Nov 2010 16:03:46 -0500 |
| parents | 63f838479510 |
| children | d840139444fe |
comparison
equal
deleted
inserted
replaced
| 605:63f838479510 | 606:bd7d11089a47 |
|---|---|
| 219 clean ones. | 219 clean ones. |
| 220 Code for generating these transformations as well as for the deep learning | 220 Code for generating these transformations as well as for the deep learning |
| 221 algorithms are made available at {\tt http://hg.assembla.com/ift6266}. | 221 algorithms are made available at {\tt http://hg.assembla.com/ift6266}. |
| 222 | 222 |
| 223 \vspace*{-3mm} | 223 \vspace*{-3mm} |
| 224 %\newpage | 224 %%\newpage |
| 225 \section{Perturbed and Transformed Character Images} | 225 \section{Perturbed and Transformed Character Images} |
| 226 \label{s:perturbations} | 226 \label{s:perturbations} |
| 227 \vspace*{-2mm} | 227 \vspace*{-2mm} |
| 228 | 228 |
| 229 \begin{minipage}[h]{\linewidth} | 229 \begin{minipage}[h]{\linewidth} |
| 256 There are two main parts in the pipeline. The first one, | 256 There are two main parts in the pipeline. The first one, |
| 257 from thickness to pinch, performs transformations. The second | 257 from thickness to pinch, performs transformations. The second |
| 258 part, from blur to contrast, adds different kinds of noise. | 258 part, from blur to contrast, adds different kinds of noise. |
| 259 \end{minipage} | 259 \end{minipage} |
| 260 | 260 |
| 261 \newpage | |
| 261 \vspace*{1mm} | 262 \vspace*{1mm} |
| 262 %\subsection{Transformations} | 263 %\subsection{Transformations} |
| 263 {\large\bf 2.1 Transformations} | 264 {\large\bf 2.1 Transformations} |
| 264 \vspace*{1mm} | 265 \vspace*{1mm} |
| 265 | 266 |
| 340 complexity]$.\\ | 341 complexity]$.\\ |
| 341 %\end{minipage} | 342 %\end{minipage} |
| 342 \end{minipage} | 343 \end{minipage} |
| 343 \vspace*{3mm} | 344 \vspace*{3mm} |
| 344 | 345 |
| 345 \iffalse | |
| 346 \vspace*{-4.5mm} | 346 \vspace*{-4.5mm} |
| 347 | 347 |
| 348 \begin{minipage}[h]{\linewidth} | 348 \begin{minipage}[h]{\linewidth} |
| 349 \begin{wrapfigure}[7]{l}{0.15\textwidth} | 349 \begin{wrapfigure}[7]{l}{0.15\textwidth} |
| 350 %\hspace*{-8mm}\begin{minipage}[b]{0.25\linewidth} | 350 %\hspace*{-8mm}\begin{minipage}[b]{0.25\linewidth} |
| 478 computed from the following element-wise operation: $\frac{image + filtered\_image | 478 computed from the following element-wise operation: $\frac{image + filtered\_image |
| 479 \times mask}{mask+1}$. | 479 \times mask}{mask+1}$. |
| 480 This module is skipped with probability 75\%. | 480 This module is skipped with probability 75\%. |
| 481 %\end{minipage} | 481 %\end{minipage} |
| 482 | 482 |
| 483 \newpage | 483 %\newpage |
| 484 | 484 |
| 485 \vspace*{-9mm} | 485 \vspace*{-9mm} |
| 486 | 486 |
| 487 %\hspace*{-3mm}\begin{minipage}[t]{0.18\linewidth} | 487 %\hspace*{-3mm}\begin{minipage}[t]{0.18\linewidth} |
| 488 %\centering | 488 %\centering |
| 601 so the image is normalized into $[\frac{1-C}{2},1-\frac{1-C}{2}]$. The | 601 so the image is normalized into $[\frac{1-C}{2},1-\frac{1-C}{2}]$. The |
| 602 polarity is inverted with probability 50\%. | 602 polarity is inverted with probability 50\%. |
| 603 %\vspace{.7cm} | 603 %\vspace{.7cm} |
| 604 \end{minipage} | 604 \end{minipage} |
| 605 \vspace{2mm} | 605 \vspace{2mm} |
| 606 | |
| 607 \fi | |
| 608 | 606 |
| 609 \iffalse | 607 \iffalse |
| 610 \begin{figure}[ht] | 608 \begin{figure}[ht] |
| 611 \centerline{\resizebox{.9\textwidth}{!}{\includegraphics{images/example_t.png}}}\\ | 609 \centerline{\resizebox{.9\textwidth}{!}{\includegraphics{images/example_t.png}}}\\ |
| 612 \caption{Illustration of the pipeline of stochastic | 610 \caption{Illustration of the pipeline of stochastic |
| 1070 with deep learning and self-taught learning. | 1068 with deep learning and self-taught learning. |
| 1071 | 1069 |
| 1072 A Flash demo of the recognizer (where both the MLP and the SDA can be compared) | 1070 A Flash demo of the recognizer (where both the MLP and the SDA can be compared) |
| 1073 can be executed on-line at {\tt http://deep.host22.com}. | 1071 can be executed on-line at {\tt http://deep.host22.com}. |
| 1074 | 1072 |
| 1075 \newpage | 1073 %\newpage |
| 1076 { | 1074 { |
| 1077 \bibliography{strings,strings-short,strings-shorter,ift6266_ml,aigaion-shorter,specials} | 1075 \bibliography{strings,strings-short,strings-shorter,ift6266_ml,aigaion-shorter,specials} |
| 1078 %\bibliographystyle{plainnat} | 1076 %\bibliographystyle{plainnat} |
| 1079 \bibliographystyle{unsrtnat} | 1077 \bibliographystyle{unsrtnat} |
| 1080 %\bibliographystyle{apalike} | 1078 %\bibliographystyle{apalike} |