Via Spreadsheet

Before getting started with the Rapid Annotator, you need first to sign up. Then, you will be assigned either an experimenter or an annotator profile. Note that only experimenter profiles can both start new projects and add annotators to them.


Create an Experiment

After logging in, you will find a Home Screen as below. Click on the “Add Experiment” button to start a new project.

The Red Hen Rapid Annotator will process data of the following types: Image, Audio, Text, and Video. Formats allowed for these categories are as the following:

  • Image: jpg, png, gif.
  • Audio: mp3, wma, mp2, wav.
  • Text: txt, html.
  • Video: mov, mp4, wmv, webm.

Add an experiment name, description, and type of data in the pop-up toolbar. Select "viaSpreadsheet" and click on “Add Experiment”. Click on “Choose Spreadsheet” and upload a spreadsheet. This must me of extension .xls (for now). Each file must be in a separate row. Do not leave any empty row in between. Start from the first row itself and fill out the following details in each respective column:

  • Name.
  • Content - for text plain text, for image / video / audio - url to the location where the actual data is.
  • Caption.

For instance, one line would look as follows:

Test image    http://placehold.it/120x120&text=image4    image for test Rapid Annotator spreadsheet tool.

Source: Placeholder.

Once files are uploaded you can further modify this info by double clicking on them in the main experiment page.

Note: Uploaded files should match type of data and allowed formats.

After you press “Start Upload”, wait till the status change and a check is displayed. Otherwise files will be not uploaded.

Hint: For audio and video experiments you will be asked for the time frame’s boundaries of the file(s) regarding the original source. This info is not mandatory. If so, set the time in seconds. Setting these boundaries will define the context of the video or audio file imported. Otherwise, it will be automatically set to 0 seconds.

Settings

Click on the “Settings” button on the main experiment screen. Here you can delete either an annotator, owner or the project itself.

Annotation

Levels of Annotation and Labels

Click on “Add Labels” in the options bar. Now click on “Add Annotation Level” to set a new category (e.g., age, gender) and press “Submit”. On the following pop-up window, you can also decide the order in which the level of annotation is asked for the data-item.

Below the level name, press “Add Label” and assign possible values for your level of annotation. For rapid annotation, you can set a Key Binding for every value belonging to an annotation level. These will be further displayed on the annotation screen. If no key is bound to any label, a key binding will be automatically made. After labels are made, close this screen by pressing the “Done” button.

Getting to Annotate

Before you begin annotating, you will need to add an annotator (at least you) among users who have already signed up. Here, you can also add new owners who will have the same experimenter privileges to add/delete data and annotators.

Now, on the Home Screen, go to “Experiments to Annotate”. You will have access to experiments you have been assigned to. By clicking on any experiment, you will see the valid labels and their key shortcuts.

On this screen, either by clicking on “female” button or pressing “f” key, your annotation is saved to the server. As soon as you select any label, the next annotation level (if any) is loaded. If the experiment consists of multiple files, the next file is rendered for annotation after the last label is displayed.

The “Undo” button reset any annotation you may wish to. However, after all files are annotated there is no a possibility to do so (for now).

The “Turn off/on looping” feature is only available for video – audio files.

Results

Go to the “Settings” button on the main experiment screen and click on “View Results”. Here you can see the results of the current project. Finally, in this screen you can either discard all annotations or export them to an .xls spreadsheet.