Skip to main content

NotebookHelpers

These are miscellaneous helper methods to render responses from ApertureDB in a jupyter environment.

This module provides helper classes (such as BoundingBox, TemporalBoundingBox) and functions (such as display, draw_bboxes, display_video_mp4, display_annotated_video) to help render visual results from ApertureDB queries directly within a Jupyter Notebook. It makes it easier to visualize bounding boxes, images, and videos without writing boilerplate code.

BoundingBox Objects​

class BoundingBox()

A class which combines a box and a label.

TemporalBoundingBox Objects​

class TemporalBoundingBox()

A class that represents a box, but also associated time in the range of frame start and end

display​

def display(images_array, save=False)

Show images with matplotlib

note

This method was used by ApertureDB to display images, but it is not recommended anymore. It will not draw annotation.

Instead, when using with JSON queries, make a instance of the Images class and call the display method. :

from aperturedb.Images import Images
result, response, blobs = execute_query(client, [{"FindImage":{"uniqueids": True}}], [])
wrapper = Images(client, response=response[0]["FindImage"]["entities"])
wrapper.display()

display_video_mp4​

def display_video_mp4(blob)

Display a video using IPython.display

Arguments:

  • blob bytearray - A blob that is the Video.

annotate_video​

def annotate_video(blob, bboxes: List[TemporalBoundingBox] = [])

Place annotations on a video framewise

Arguments:

  • blob bytearray - Video blob returned from the database
  • bboxes List[TemporalBoundingBox], optional - List of boxes to be drawn. Defaults to [].

display_annotated_video​

def display_annotated_video(blob, bboxes: List[TemporalBoundingBox] = [])

Returns a HTML representation with a column filled with video entities.