deletions | additions       diff --git a/figures/clustering/caption.md b/figures/clustering/caption.md  index aa90798..b79458c 100644  --- a/figures/clustering/caption.md  +++ b/figures/clustering/caption.md 
...
\label{fit:cluster}  **Classifying detected events.** **A** Based on the 4 shape parameters (see Figure \ref{fig:fit}), a density Density  based clustering algorithm is used to classify pixel events into categories. Plots show projections of 4-dimensional parameter space. categories based on their shape.  Colours represent event categories. Black signifies events which could not be categorized. **B**  Categories obtained from clustering by shape plotted according to event location. **B** **C**  In the second step of classification each shape category is clustered further based on location. Groups of pixel events are obtained that make up a Ca2+ release even (in this case sparks and a wave). **C** Using only pixel events that were successfully classified **D** All detected sparks visualized according to their center(location of coloured circle) and amplitude (colour of the circle). Events marked with  a new ΔF/F0 image can be built using only black x were not symmetric and not  classified events (compare as sparks. Circle colour shows sparks originating from the same 2um segment. Empty circles are sparks  with \ref{fig:linescan}E) no nearby sparks.