Skip to content
Permalink
Browse files

Fleshed out index

  • Loading branch information...
danlester committed May 29, 2019
1 parent 8fe5ef6 commit 141cb0c8f8be21e344b0913530c3b4e4a8b0e0ec
Showing with 54 additions and 52 deletions.
  1. +2 −0 docs/chapters/installation.rst
  2. +36 −0 docs/chapters/overview.rst
  3. +1 −1 docs/conf.py
  4. +15 −51 docs/index.rst
@@ -1,3 +1,5 @@
.. _installation:

Installation
------------

@@ -0,0 +1,36 @@
.. _overview:

Overview
---------

In a data science or machine learning project, you may prepare and study
images or other data within a Jupyter notebook then need to annotate the
data to augment the training or fix errors in your source data. You may
need to check classifications are correct (cats/dogs correctly
specified) or add bounding boxes around the pertinent parts of your
images. For example, to build a dog breed classified you might first
build a model that learns how to identify the bounding box of a dog
within the image, then your final model zooms in on that box in order to
train/evaluate the breed classifier.

Since you are already working within a Jupyter notebook, the Innotater
works inline allowing you to interact with your data and annotate it
quickly and easily, syncing straight back to your input data arrays or
matrices.

Within Jupyter, you can easily home in on problem input data - perhaps
only misclassified images - so you can step through and adjust bounding
boxes just for those items.

The Innotater widget is designed with a flexible API making it quick and
easy to get started exploring your data, guessing how to work with your
data without explicit configuration where possible.

The widget is currently in ALPHA development phase, and I appreciate all
feedback on any problems including details on how the current code works
or fails to work for the structure of your particular projects.

.. figure:: ../_static/screenshots/ImageAndBBoxesFull.png
:alt: Screenshot of Innotater widget in Jupyter

Screenshot of Innotater widget in Jupyter
@@ -17,7 +17,7 @@

# -- Project information -----------------------------------------------------

project = 'innotater'
project = 'jupyter-innotater'
copyright = '2019, Ideonate'
author = 'Ideonate'

@@ -1,20 +1,28 @@
.. innotater documentation master file, created by
sphinx-quickstart on Wed May 29 11:24:36 2019.
You can adapt this file completely to your liking, but it should at least
contain the root `toctree` directive.
Welcome to innotater's documentation!
=====================================
Welcome to Jupyter-Innotater's documentation!
=============================================

Annotate data including image bounding boxes inline within your `Jupyter
notebook <https://jupyter.org/>`__ in python.

Innotater is a Jupyter notebook extension (widget) for machine learning practitioners. Use Python code to load your
training and test data, then interactively annotate the data by drawing bounding boxes, selecting
classes from a drop-down, or through other Javascript widgets.

Read a full description in :ref:`overview` or try out example Jupyter notebooks in Binder:

|Binder|

Easy local :ref:`installation` using pip.


.. toctree::
:maxdepth: 2
:caption: Contents:
:caption: Documentation Contents:

chapters/overview
chapters/examples
chapters/usage
chapters/installation
@@ -23,53 +31,9 @@ notebook <https://jupyter.org/>`__ in python.
chapters/license



Indices and tables
==================

* :ref:`genindex`
* :ref:`modindex`
* :ref:`search`





1 - Overview
------------

In a data science or machine learning project, you may prepare and study
images or other data within a Jupyter notebook then need to annotate the
data to augment the training or fix errors in your source data. You may
need to check classifications are correct (cats/dogs correctly
specified) or add bounding boxes around the pertinent parts of your
images. For example, to build a dog breed classified you might first
build a model that learns how to identify the bounding box of a dog
within the image, then your final model zooms in on that box in order to
train/evaluate the breed classifier.

Since you are already working within a Jupyter notebook, the Innotater
works inline allowing you to interact with your data and annotate it
quickly and easily, syncing straight back to your input data arrays or
matrices.

Within Jupyter, you can easily home in on problem input data - perhaps
only misclassified images - so you can step through and adjust bounding
boxes just for those items.

The Innotater widget is designed with a flexible API making it quick and
easy to get started exploring your data, guessing how to work with your
data without explicit configuration where possible.

The widget is currently in ALPHA development phase, and I appreciate all
feedback on any problems including details on how the current code works
or fails to work for the structure of your particular projects.

.. figure:: _static/screenshots/ImageAndBBoxesFull.png
:alt: Screenshot of Innotater widget in Jupyter

Screenshot of Innotater widget in Jupyter
Innotater source code can be found on `GitHub here <https://github.com/ideonate/jupyter-innotater>`_.


.. |Binder| image:: https://mybinder.org/badge_logo.svg
:target: https://mybinder.org/v2/gh/ideonate/jupyter-innotater/master?filepath=Example%2FExamples.ipynb

0 comments on commit 141cb0c

Please sign in to comment.
You can’t perform that action at this time.