What is this?


Extract Bag-of-Features histogram.

For other information about object recognition using Bag-of-Features, see here.

Subscribing Topic

  • ~input (posedetection_msgs/Feature0D)

    Feature data.

  • ~input/label (sensor_msgs/Image)

    Label image with witch features will be extracted.

Publishing Topic

  • ~output (jsk_recognition_msgs/VectorArray)

    Extracted feature data.


  • ~bof_data (String, required)

    Path to Bag-of-Features data.

    The file contains jsk_recognition_utils.feature.BagOfFeatures object, and it should be pickled and gzipped.

  • ~queue_size (Int, default: 10)

    Maximum number of messages which subscriber keeps in the queue.

  • ~approximate_sync (Bool, default: False)

    Whether to use ApproximateSync for input topics.

  • ~slop (Float, default: 0.1)

    Number of seconds you allow about the difference of timestamp.

    This parameter is enabled only when ~approximate_sync is True.


roslaunch jsk_perception sample_bof_object_recognition.launch