Skip to content
Loic Royer edited this page Oct 21, 2015 · 57 revisions

Welcome to the ClearVolume Wiki

ClearVolume Logo

ClearVolume is a real-time live 3D visualization library designed for high-end volumetric microscopes such as SPIM and DLSM microscopes. With ClearVolume you can see live on your screen the stacks acquired by your microscope instead of waiting for offline post-processing to give you an intuitive and comprehensive view on your data. The biologists can immediately decide whether a sample is worth imaging.

ClearVolume can easily be integrated into existing Java, C/C++, Python, or LabVIEW based microscope software. It has a dedicated interface to MicroManager/OpenSpim/OpenSpin control software. ClearVolume supports multi-channels, live 3D data streaming from remote mircroscopes, and uses a multi-pass Fibonacci rendering algorithm that can handle large volumes. Moreover, ClearVolume is integrated into the [Fiji/ImageJ2](Installing Fiji Plugin) and [KNIME](Installing KNIME Plugin) ecosystems - you can now open your stacks with ClearVolume from within these popular frameworks for offline viewing.


Prerequisites

  • Java JDK >= version 7 (1.7) (get it here)
  • Gradle 2.1 (get it here)
  • Requirements for either the CUDA or OpenCL backend
  • A graphics card with at the very minimum 500M of GPU RAM, preferably >=1G.

Windows specific prerequisites

  • On windows you need VisualStudio installed (get it here) to be able to use the CUDA backend and to build ClearVolume. If you use an 'express' version of VisualStudio, you need to copy and rename the file 'vcvarsx86_amd64.bat' located in folder 'Microsoft Visual Studio 12.0\VC\bin\x86_amd64' to 'vcvars64.bat' that should be placed at the root of the VisualStudio folder e.g. 'C:\Program Files (x86)\Microsoft Visual Studio 12.0'. run gradle test an check for messages that suggest that that file could not be found.

Mac OSX specific prerequisites:

  • Java 8 installed. Please note that older version of java installed are known to cause problems. If you face difficulties starting the client try to remove older version of Java.
  • At least Mac OSX 10.9 to be able to run the ClearVolume client

Linux specific prerequisites:

  • A recent version of Linux, tested: Ubuntu 14+ , Xubuntu 14+. Follow instructions here for installing CUDA.
  • Note: we have currently difficulties with a weird bug on Linux, ClearVolume works from Eclipse but not from the command line.

CUDA backend

  • NVidia graphics card
  • CUDA SDK 7.5 (get it here)

OpenCL backend

  • OpenCL 1.2 capable graphics card with preferably at least 1G of GPU RAM.

####LabVIEW bindings

  • LabVIEW 2012 64 bit.

Note: LabVIEW users can regenerate the LabVIEW bindings from almost any LabVIEW 64bit version using the LabVIEW wizard for interfacing with native C/C++ libraries. Unfortunately 32 bit is out of the questions in the 21st century - and makes little sense when handling volume data anyway...

Installing

  1. Get the code by either checking out the project using Git

     git clone https://github.com/ClearVolume/ClearVolume.git
    

    or go to the download page and download the whole repository.

    This download will only provide you with the source code. To obtain usable binaries, you need to follow the instructions below or have a look at the other downloads provided (for C/C++ and LabVIEW users).

  2. Go to the project folder root and run (that steps needs Gradle):

    Linux/OSX:

     ./build.sh
    

    Windows:

     build.bat
    

    This will generate a jar executable here:

     ./build/executable/ClearVolume.exe.jar
    

    (It uses the magic Capsule plugin for Gradle for that)

Simple demo

Accessing our demo dataset over the network

To demonstrate the remote rendering capabilities, there is a data server running on

clearvolume.mpi-cbg.de

which you can use for a quick demonstration.

First, start the network client (OSX):

./build/executable/ClearVolume.exe.jar

If the above does not work or if you are on Linux or Windows you might have to type instead:

java -jar ClearVolume.exe.jar

You can also simply double click on the ClearVolume.exe.jar file (OSX)

Now point the Client to the url mentioned above (that should be the default):

Smaller icon

and you should see some nice mitotic divisions in a C. Elegans embryo:

Smaller icon

You can adjust min, max, and gamma values with the Keyboard/Mouse shortcuts if necessary). You can also press the key 'C' while the rendering window has focus which will bring up a control window (which can be closed by pressing 'C' again).

start a demo server locally

Start the demo server (Linux or OSX):

./build/executable/ClearVolume.exe.jar --demo-server > log.txt &

Start the network client (Linux or OSX):

./build/executable/ClearVolume.exe.jar

In windows you might have to type instead:

java -jar ClearVolume.exe.jar

You can also simply double click on the ClearVolume.exe.jar file (OSX, Linux)

Connect to the server on localhost

Client

You should see a volume rendering of a 3D Volume stream of changing 3D data (adjust the gamma value with the Keyboard shortcuts if necessary).

ClearVolume Demo

ClearVolume network client

The ClearVolume network client allows you to connect to your ClearVolume enabled microscope software from anywhere in the planet where your internet connection is fast enough. In the worst case scenario it just takes a bit of time to download the data... Since the rendering is happening on the client, once you have a stack/volume on your client computer you can rotate and zoom virtually all you want. The ClearVolume network client (ClearVolume.exe.jar) is also available for download here

Smaller icon

You need to provide a network address, decide on a window size. Importantly you need to specify the type of data: typically 16 bit and 1 color. If you provide the wrong information the client will return an error. We will try to improve this in a future version. If you want to use the time-shift and channel filter features just tick the mark. Note: remote time-shift only works with stacks received by the client.

How do I integrate ClearVolume into my control software?

One of ClearVolume key feature is to make it easy and painless to integrate into different systems. For this purpose we have made sure that both the APIs are simple and elegant as well as to provide bindings for all languages that we know are used for microscope control, and in particular light-sheet microscopes.

ClearVolume comes with bindings with all major languages used for developing microscope control software: Python, Java, C/C++, LabVIEW. There are two types possibilities for integration:

  1. Network : ClearVolume receives data over the network via streaming
  2. In-Process: ClearVolume lives in the same process and thus data can be transferred at maximal speed.

It is possible to have both: in-process with ClearVolume server listening for incoming connections, this mode offers the possibility to monitor long-term time-lapses remotely.

Different integration guides can be found here:

OpenSPIM and ClearVolume

OpenSPIM has democratised light-sheet microscopy and has made it possible for many labs around the world to build high-quality and cheap light-sheet microscopes. To support this effort we integrated ClearVolume into the OpenSPIM plugin for μManager. Go to this page OpenSPIM Manual for details on how you can use ClearVolume with OpenSPIM. If you are interested in how this integration was done, have a look at Integration in OpenSPIM.

Contribution guidelines

  • torture of animals with the exception of bugs is prohibited
  • exhibit openness of mind towards idiosyncratic ideas
  • be nice

Contributors

  • Loic Royer (royer -at- mpi-cbg -point- de)
  • Martin Weigert (mweigert -at- mpi-cbg -point- de)
  • Ulrik Guenther (guenther -at- mpi-cbg -point- de)
  • Florian Jug (jug -at- mpi-cbg -point- de)