One method, which may not necessarily correlate to a boxplot in all situations (but probably works for the default boxplot), is to use isoutlier:
T = [1;-1;0.5;-0.5;0.75;-0.75;-6;-7;4;3];
numOutliers = sum(outliers)
h = boxplot(T,'Labels',{'1'});
If you need to be sure you have the exact count in the boxplot, you can get the data in the image and drill down to find the outliers:
plts = bxplt.Children
plts =
7×1
Line array:
Line (Outliers)
Line (Median)
Line (Box)
Line (Lower Adjacent Value)
Line (Upper Adjacent Value)
Line (Lower Whisker)
Line (Upper Whisker)
numRedMarks = numel(plts(1).YData)
To programmatically find the plots of outliers (in the event that you have more than one box):
idxs = find(strcmp({plts.Tag},'Outliers'));
outlierCell = {plts(idxs).YData}
numMarksFirstPlot = numel(outlierCell{1})