What is this?

Face recognition using Amazon Rekognition, see for more info.

Subscribing Topic

  • ~image (sensor_msgs/Image)

    Raw image.

  • ~face_roi (opencv_apps/FaceArrayStamped)

    Rectangles on the face of input image. Use ROI value.

msg.faces[].face.x      : X coordinates of the center of the face image in the ~image input
msg.faces[].face.y      : Y coordinates of the center of the face image in the ~image input
msg.faces[].face.width  : Width of the face image
msg.faces[].face.height : Height of the face iamge

Publishing Topic

  • ~face_name (opencv_apps/FaceArrayStamped)

    Publish recognized face name as well as face image. The face.{x,y,width,height} corresponds to input face_roi, that means x, y is the center of face rectangle.


  • ~use_window (Bool, default: False)

    Show input image on the window, if it is true.

  • ~env_path (String, default: env.json)

    Json file for environment variables to run aws auto-checkin app. You can find how to generate this file on In addition to that, you need to add “UserName” and “UserPassword”

  "Region": "%%REGION%%",
  "ApiEndpoint" : "",
  "CognitoUserPoolId": "%%COGNITO_USER_POOL_ID%%",
  "CognitoUserPoolClientId": "%%COGNITO_USER_POOL_CLIENT_ID%%",
  "FaceAreaThreshold": 1e4,
  "FaceMarginRatio": 0.2,
  "FaceSimilarityThreshold": 90,
  "CroppedImageWidth": 540,
  "CroppedImageHeight": 540,
  "NameTtlSec": 10,
  "UseDeepLeaningForDetector": true,
  "UserName": "%%YOUR_USER_NAME%%",
  "UserPassword": "%%YOUR_PASSWORD%%"


roslaunch jsk_perception sample_aws_auto_checkin_app.launch use_window:=true

For JSK user, Download env.json file from Gdrive and put this under /tmp directory to run sample code.

To add new people to face database, add face image file to Amazon S3, auto-check-in-gapp-register… buckets