Tools for evaluating the performance of the Waggle Dance Detector (WDD) system.
conda activate beesbook
pip install git+https://github.com/BioroboticsLab/bb_wdd_eval.git
annotation_matcher.py
: Given a spreadsheet of manually annotated waggle phases, will identify candidate matches in the automatically annotated WDD data. Matching is based on timestamp and spatial proximity within configurable tolerance thresholds.
coordinate_transformer.py
: Contains a class for transforming coordinates between HD camera reference frame (used in manual annotations) and the WDD camera reference frame (used in automatic annotations).
python3 annotation_matcher.py --help