Occiput

What is occiput?
Occiput is a computing platform for volumetric imaging. Occiput provides a Python library for image reconstruction and processing and a web-based interface (based on IPython and TipiX) which enables the use of occiput on the cloud.
The core Python library of occiput provides tools that enable the development of software for image formation and processing. Currently occiput is specialized in tomographic reconstruction, with tools for Single Photon Emission Computed Tomography (SPECT), Positron Emission Tomography (PET) and X-Ray Computed Tomography (though this last one is incomplete) and in volumetric affine image registration. The design of occiput unifies tomographic reconstruction and registration, providing a framework for the development of algorithms for dynamic tomographic imaging.
Occiput is optimized for Graphics Processing Units (GPU), enabling the development of near real-time reconstruction algorithms.

Run in the cloud
An instance of occiput can be found here. This can be used to test occiput without installing it on a computer.

Install
First, install NiftyRec (NOTE: if built manually, it is not necessary to enable the Python extension when configuring with CMake). You are nearly done, now install you need to install occiput and the Python packages that it depends on. There are two ways:
1) Using PIP (the Python Package Index) – the dependencies are installed automatically. At the command line (under MacOS, Windows or Linux), type:

pip install occiput

PIP should download and install all the dependencies. PIP may fail to install nibabel or scipy (two of the dependencies of occiput). If this happens, sorry, not our fault. Find a solution with Google, then repeat pip install occiput.

2) Building the Python package from source.
Download the occiput source code from Github and install manually. (Note however that you must also install the dependencies, via PIP or manually. Most of the dependencies of occiput are hosted on occiput.io).

python setup build test install 

Get started
occiput is currently mostly undocumented. The best way to get started with occiput is to install IPython Notebook and try the notebooks that come with occiput in the /notebooks folder:

cd notebooks 
ipython notebook  

The notebooks provide examples of PET reconstruction, SPECT reconstruction, loading and saving files, interactive image visualization, volume rendering, rigid volumetric image registration, loading data from Dropbox and Google Drive.