this is for holding javascript data
Paul St-Aubin edited Results.tex
over 9 years ago
Commit id: 711e93cbf810071bb2df8d43eb8670755287909e
deletions | additions
diff --git a/Results.tex b/Results.tex
index f537be1..a482a68 100644
--- a/Results.tex
+++ b/Results.tex
...
\section{Experimental Results}
A set of video data \subsection{Tracking calibration}
\label{tracking_calibration}
Tracking accuracy was
collected at 20 roundabouts, 40 merging zones evaluated by way of
varying lane configuration, geometry, land use, and traffic volumes across Measures Of Tracking Accuracy (MOTA) \cite{ettehadieh15systematic} using as ground truth manually annotated trajectories with the
province annotation tool \href{http://www.jpjodoin.com/urbantracker/index.html}{Urban Tracker} \cite{Jodoin_2014}. Nearly 40,000 manually annotated observations of
Qu{\'{e}}be. The merging zone 371 motor vehicles were performed across two prototypical sample sites: one site where objects were tracked at a distance of
between 20 and 50 meters away from the
roundabout is defined as camera and one site where objects were tracked at a distance no greater than 20 meters away. An initial MOTA was calculated for default tracking parameters and offered modest accuracy (in the
portion neighbourhood of
$70\%$). A tracking optimisation was then performed using a genetic algorythm to search for parameters that maximised MOTA. After optimisation, accuracy increased up to $94\%$. These statistics are summarised in Table~\ref{tab:track_calib_data}. Tracking parameter optimisation converged in less than 24 hours.
Of important note is the
ring intersected by an approach and an exit. Video data at each site fact that this calibration was
taken performed on
an average summer workday each from 6 AM the oldest and poorest quality video data (resolution of $800\times 600$, without software-assisted image stabalisation, and before lens correction). Tracking results are expected to
7 PM or 10 PM increase with increases in quality of the video data and
captures both peak traffic hours. \cite{St_Aubin_2013} video pre-treatment. The optimized tracking parameters should be portable to other sites with similar view and camera characteristics, though this will be investigated further and the subject of of a future paper in depth.
\begin{table}
\caption{Tracking Calibration Results}
\label{tab:track_calib_data}
\begin{tabular}{lllll}
\hline
\textbf{Site} & \textbf{Ground Truth Objects} & \textbf{Ground Truth Observations} & \textbf{Unoptimised MOTA (Default Parameters)} & \textbf{Optimised MOTA} \\
\hline
12 (far camera) & 146 & 12887 & 0.685 & 0.944 \\
16 (near camera) & 225 & 25011 & 0.744 & 0.853 \\
\end{tabular}
\end{table}
\subsection{Data Size}