How to join as an annotator?
An annotator annotates the vocalizations from the recordings of a dataset by following the guidelines set by the corresponding investigator(s).
Here is the expected timeline and key steps in the process:
Step 1: Sign the VoCallBase Collaboration and User Agreement as a User and upload it via the signup form
Step 2: VoCallBase admin team approves your signup and sends you further instructions together with an access token. This access token is needed for all interactions with VoCallBase resources.
Step 3: Carefully read the annotation guidelines of the dataset as provided at Step 2
Step 4: Practice with artificial examples to get familiar with CallMark before proceeding
Step 5: Use the access token to sign in to our annotation platform to annotate voice activity (VA) segments in the chosen dataset. The expected time is about 8 hours for task completion.
Now benefit from WhisperSeg!
You are now free to use WhisperSeg and the entire human-in-the-loop system!
The submitted annotations are now used to train and fine-tune the WhisperSeg model. Using your unique access token, you will gain unrestricted access to the WhisperSeg tool on CallMark, enabling them to automatically annotate the entire dataset.
Now you can correct the annotations, submit them again, and teach WhisperSeg even more about your dataset and species!