deletions | additions       diff --git a/DeclareMathOperator_argmin_arg_min_DeclareMathOperator__1.tex b/DeclareMathOperator_argmin_arg_min_DeclareMathOperator__1.tex  index 47f2852..0a1df1a 100644  --- a/DeclareMathOperator_argmin_arg_min_DeclareMathOperator__1.tex  +++ b/DeclareMathOperator_argmin_arg_min_DeclareMathOperator__1.tex 
...
used. Therefore, the training dataset will not show any variance, if  the same images were used.  \subsection{Texton Dictionary Generation}  \label{sec:text-dict-gener}  For generating a suitable dictionary for a given environment, an  initial flight was performed. During this flight, 1000 randomly  selected image patches of size $w \times h = 6 \times 6$ were  extracted per image. In total, 100 images were used, resulting in  $100,000$ image patches, which were clustered to create a texton  dictionary. The resulting cluster centers---the prototypes of the  clustering result---are the textons~\cite{varma2003texture}. An  example of a learned dictionary can be found in  Figure~\ref{fig:dictionary}.  Different situations require different textons and a different number  of them. The choice of these parameters is map-dependent, and we set  it to 20 textons for all maps.  % TODO: Write some stuff about the low-level implementation