Redaction Assistant

Redaction Assistant automatically detects and masks common objects such as heads, license plates, non-phone screens (MDTs, MDCs, computer monitors, television screens). This feature is an Axon Redaction Studio add-on, and might not be available in all agencies.

Watch this video to learn how to use Redaction Assistant.

Guided Redaction Assistant process

  1. Select the piece of video evidence you wish to redact.
  2. Select the Redactions tab to launch Redaction Studio within your browser window.
  3. Select the Assistants tab.
    Redaction pane with teh Assistants tab highlighted.
  4. Choose a Redaction Assistant process for your video by selecting Start next to one of the following options:
    • Notebooks: This process checks the video for physical notebooks or notepads that an officer might write in while at a scene. This also includes things like vehicle registration or other paper items.
    • Screens: This process checks the video for non-phone screens and adds masks to cover the screens.
    • Mobile devices: This process checks the video for cellphone screens that are on (illuminated) and adds masks to cover the screens.
    • License plates: This process checks the video for vehicle license plates and adds masks to cover the plate.
    • Heads: This process checks the video for heads and adds masks to cover the heads, and then provides users with the option to choose which individual heads to redact.
    • ID documents: This process checks the video for things like drivers licenses and passports and adds masks to cover the objects.
    • Audio PII: This process detects personally identifiable information (PII) in an audio transcript so you can review detected keywords and apply audio redactions. Redactions are applied to the audio of the video and appear in the waveform. To change what Audio PII detects or how it redacts, update your settings (Assistant settings > Audio Redaction).
  5. Monitor the progress bar to track the status of the redaction.
    While Redaction Assistant is processing, you can:
    • Continue redaction work on the video.
    • Adjust Redaction Assistant masks.
    • Begin new redaction tasks.
    • Exit Redaction Studio and Redaction Assistant will continue to scan and process the video.
  6. After the scanning is complete, a green checkmark displays next to the assistant you ran. The number of detected items is shown next to the assistant type. To review detection details, select the chevron (>) next to the item count.
    Redaction pane showing a completed ID documents scan with a green checkmark next to it.
  7. Select a redaction from the list to review it.
  8. Select the mask in the video segment or timeline to review its placement and duration. Watch the entire video to verify masks were added to the necessary items.
    Mask selected on the video timeline showing the Redact option on the mask in the video.

    Note

    The redaction behavior of masks is determined by the Redaction Assistant settings. To include or exclude an object from redaction, check or uncheck the Redact box. For more information on configuring these settings, see Redaction Assistant Settings.

  9. Select and then review each mask in the segment timeline or Redacted Objects tab. While reviewing the masks, you can:

    • Adjust the position or size of the mask as needed.
    • Change the blur level.
    • Delete a mask by selecting the Delete button, or delete all masks of a certain type by selecting X on the corresponding process bar.
    • Adjust the start and end times for the duration the mask appears.
    • Use Spray Paint Redaction masks to extend segments with finer control. Select and hold on the mask during video playback at normal, half speed, or rewind, and then use your mouse to follow the object you wish to redact.
  10. After you have reviewed and edited the Redaction Assistant masks and added any of your own masks, select Extract to extract the video.

Redaction Assistant processes up to 15 masks for faces at a time. If more than 15 faces are detected, the first 15 are processed and the remaining masks are placed in a queue to wait their turn. If it appears that Redaction Assistant isn’t processing all face masks, wait for processing to complete for all the selected objects before adding your own masks.

Bulk redaction

Bulk Redaction shows all identified faces in a thumbnail view so you can see all results on a single screen, allowing you to quickly select which faces to redact throughout an entire video.

  1. Search for the evidence files you wish to include in the bulk redaction.
  2. Select the checkbox next to each file you want to redact.
  3. Select More actions on the action bar, then select Redact.
    Bulk action bar with the More actions menu expanded.
  4. In the bulk request pane, enter a Title and Description for the bulk redaction.
  5. Choose which objects to detect and redact. Choose from:
    • Notebooks
    • Screens
    • Mobile devices
    • License plates
    • Heads
    • ID documents
    • Audio PII
      Bulk redaction options dialog box.
  6. Select View Settings to set the redaction options and blur level for each object type.
  7. Select Create and Run.
  8. In the confirmation window, select View Request to view the Redaction Assistant status.
  9. Select Redact next to each item to open it in Redaction Studio.

Audio PII bulk prerequisites

You can start the Audio PII assistant from bulk redaction even if a transcript is not already available.

If your organization does not have auto transcription enabled, generating a transcript uses transcription minutes.

If you start the Audio PII assistant from within Redaction Studio, generate a transcript from the Transcript tab before you run the Audio PII assistant.

Redaction Assistant settings

The Redaction Assistant settings allow you to configure your own default settings for Redaction Assistant masks. These settings can be changed before starting Redaction Assistant, during the review process, or after finishing the redaction, offering both flexibility and efficiency.

Manage your AI assistant mask settings from the Assistants tab in Redaction Studio.

  1. Select Assistant settings (gear).
  2. To configure video detection and redaction, select the Visual Redaction tab.
    Assistant settings with the Visual redaction tab highlighted.
  3. For each object type, choose whether you wan the assistant to Auto-redact or Detect only.
  4. Select Save.
  5. To configure audio detection and redaction, select the Audio redaction tab.
    Assistant settings with the Audio redaction tab highlighed.
  6. Under Audio objects, select the audio PII types you want the assistant to detect.
  7. For each selected audio object, choose whether you want the assistant to Auto-Redact or Detect only.
  8. For audio objects you set to auto-redact, choose an audio mask type (Mute or Short bleep).
  9. Select Redaction margins, then choose how much time to add before and after detected speech so sensitive words are fully redacted. These settings apply when redacting audio from a transcript:
    • None Short (250 ms)
    • Medium (500 ms)
    • Long (1000 ms)
    • Custom
      Assistant settings with the redaction margin options highlighted.
  10. Select Save.