Experiments and Results

First the performance of the classifier is shown. Then, the attention algorithm is additionally applied to adapt the detection and to reduce the false positives.