Dtcwt

The big news in this release is a new TensorFlow backend contributed by fbcotter.

This release adds no significant functionality over 0.10.0 but includes fixes for certain nVdidia OpenCL configurations and some corner-cases when running dtcwt under Python 2. It also updates the documentation to be more DFSG-friendly.

dtcwt is a Dual-Tree Complex Wavelet Transform library for Python along with a collection of DT-CWT algorithms and GPU acceleration.

As always full documentation is available via [read the docs](https://dtcwt.readthedocs.org/).

This release includes significant speedups to the image registration algorithm along with some more bug fixes for the OpenCL pipeline.

dtcwt is a Dual-Tree Complex Wavelet Transform library for Python along with a collection of DT-CWT algorithms and GPU acceleration.

As always full documentation is available via [read the docs](https://dtcwt.readthedocs.org/en/0.9.0/).

This is a simple bugfix release which addresses a single bug in the OpenCL backend whereby if a user passed an explicit OpenCL context to be used, the default context would be used instead in some cases.