ibb 2.1.0

Jupyter Widget Library for Brambox

Interactive BramBox

IPython widgets for visualizing brambox annotation and detection dataframes on their images. These widgets help with doing computer vision research in IPython Notebooks.

Widget List

You can find more information about each widget by checking its doc-string:

import ibb

# Python
help(ibb.<WIDGET_NAME>)

# Alternative in Notebook
?ibb.<WIDGET_NAME>

• ImageCanvas: Barebones widget to draw numpy arrays as images and rectangles. Use this to create your own widgets.
• ImageViewer: Image list browser.
• BramboxViewer: Brambox dataset browser. Browse through your annotations or detections for visual inspection.
• PatchViewer: Brambox dataset browser, but cut images in smaller patches (with overlap) for easier visualisation.
• CutoutViewer: Brambox dataset browser, but view individual annotations or detections as cutouts of the image.

Installation

To install use pip:

> pip install ibb
> jupyter nbextension enable --py --sys-prefix ibb

For a development installation (requires npm),

> git clone https://github.com/0phoff/ibb.git
> cd ibb 
> pip install -e .
> jupyter nbextension install --py --symlink --sys-prefix ibb
> jupyter nbextension enable --py --sys-prefix ibb