From 78fb80ec68a006b3879ccd3f0d19d182db320878 Mon Sep 17 00:00:00 2001 From: Jim Martens Date: Sun, 15 Sep 2019 13:50:28 +0200 Subject: [PATCH] Added first bit of qualitative analysis Signed-off-by: Jim Martens --- body.tex | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/body.tex b/body.tex index 96b45df..330bda2 100644 --- a/body.tex +++ b/body.tex @@ -843,6 +843,26 @@ are included. All plotted variants show a similar behaviour that is in line with previously reported figures, such as the ones in Miller et al.~\cite{Miller2018} +\subsection*{Qualitative Analysis} + +% TODO: expand + +This subsection focuses not on the big picture but compares vanilla SSD +with Bayesian SSD with respect to specific images that illustrate +similarities and differences between both approaches. For this +comparison, 0.2 confidence threshold is applied. Furthermore, Bayesian +SSD uses non-maximum suppression and dropout with 0.9 keep ratio. + +\begin{figure}[ht] + \centering + \includegraphics[width=0.5\textwidth]{COCO_val2014_000000336587} + \caption{Image with stop sign and a truck at the right edge. Origin of the image is the top left corner.} + \label{fig:stop-sign-truck} +\end{figure} + +In the ground truth, only a stop sign and a truck are detected in figure \ref{fig:stop-sign-truck}. The truck is neither detected by vanilla nor Bayesian SSD. Instead both detected a pottet plant and a traffic light. +The stop sign is detected by both variants. + \chapter{Discussion and Outlook} \label{chap:discussion}