Skip to content

Annotation

Today DASexplorer supports four annotation types, each suited to different detection or labeling tasks. Each type is saved to its own CSV file, distinguished by a suffix appended to the data file's stem:

Type Use case Docs File suffix
BBox Axis-aligned bounding boxes (e.g. object detection) BBox _bbox.csv
OBBox Oriented bounding boxes (e.g. angled events) OBBox _obb.csv
Keypoints Point-based labeling (e.g. call onsets) Keypoints _kp.csv
Line Line-based labeling (e.g. tracked trajectories) Line _lin.csv

Note

For example, annotating myfile.h5 produces myfile_bbox.csv, myfile_obb.csv, myfile_kp.csv, and/or myfile_lin.csv, depending on which annotation types were used. This keeps annotations of different types separated and avoids naming collisions.

Creating annotations

To annotate events, go to the Annotate section of the left panel and click one of the annotation types.

Annotation panel

A crosshair guide following the pointer indicates that you are in edit mode. Draw the annotation on the visualization, then define an ID and a comment for it.

Tip

Once an annotation is created, right-click it to bring up two options: Edit Event to modify its ID and comment, and Edit Shape to adjust its geometry.

To delete an existing annotation, right-click it and select Remove Event.

Managing annotations

The Annotations section of the left panel lists all annotations currently present in the session.

From this list you can:

  • Save CSV — export the current annotations to a CSV file
  • Clear All — remove all current annotations

At the bottom of this section, the path where annotation files are saved is displayed. By default, this is the same directory as the data file, but you can change it at any time with the ... button. If previous annotation files are found in this path, they will be displayed automatically. By convention, for each type of annotation (e.g., BBox), one CSV file is saved for each annotated DAS file.

Next steps