- Override random colors with your own
- Override image captions with your own
- Choose to show or not show masks
- Choose to show or not show bounding boxes

If you stop training at epoch 15 then continue
from where you left off, it resumes with 17 which
skips 16. Does cause any problems other that being
a bit annoying.

Before the fix, the computed AP score is slightly
lower than it should be. It doesn’t affect
training, but just shows slightly lower results AP
in the notebooks.

This is useful to run prediction on images of
variable sizes without resizing them to a fixed
size. While using mode=“none” is another option,
it requires that you pre-process your images to
ensure the width and height are multiples of 64.
This mode handles that automatically.

- Update ROOT_DIR depending on path of file
- Remove unneeded __init__.py files
- Remove setting matplotlib backend

* create package mrcnn
* move notebooks to samples

A step towards supporting detection on different
image sizes.

This also changes utils.resize_image() such that
it doesn’t convert images to 0-255 range, but
rather keep the input range. This would only
affect users who use input image of range 0-1
rather than 0-255.

This doesn’t change the functionality, other than
cleaning up the code to prepare for future changes
to support detection on different image sizes.

In a previous commit, the detections tensor was
changed to use normalized coordinates [0, 1]. This
fixes the notebooks to visualize the updated
detections format correctly.

The image_meta tensor stores meta data about the
image. This update adds the scale by which the
image was resized and it’s final shape.

Make DetectionLayer return normalized coordinates
to avoid unnecessary conversion to pixels and back
to normalized.

Move code to separate functions and fix off-by-1
error.

Allows using a config variable, TRAIN_NB, rather
that updating the BatchNorm class.

skimage.transform.resize() crashes with a floating
point exception if the image has 1-pixel width or
height and the mode=“reflection”. Changing mode to
“constant” (the default) fixes the issue. And, 
it’s the correct mode to use anyway.

If the mask is a rectangle, its mini-mask pixels
become all 1s, and it has no zeros. Scipy imresize
fails on that and returns 0s instead. This update
fixes the problem.

Requires installing imgaug (pip3 install imgaug)
https://github.com/aleju/imgaug

List of augmentations:
http://imgaug.readthedocs.io/en/latest/source/augmenters.html

This is more reliable than using the batch size to
determine worker count because load per batch of 
images varies depending on image sizes and where
they’re loaded from.