ICON is an interactive tool for training deep neural networks for image segmentation tasks. A user enters sparse annotations over a web-based user interface to train a classifier running on a high-performance GPU-enabled server. The classifier produces pixel confidences that are rendered as an overlay on the user interface to guide the the annotation process. The server needs to be setup only once on a single machine or a cluster; and the end users require a browser (Chrome or Firefox) to access the system.
MIT License
Copyright (C) 2016 Harvard University
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
The UI runs on a web browser while the classifiers runs on a server with GPU
cython h5py hdf5 jpeg keras libpng libtiff mahotas matplotlib numpy opencv pandas pil pillow scikit-image scikit-learn scipy sqlite theano tornado
-
Run install.sh once, to setup the system (This should be done on a linux system)
-
Start the web server by running: sh web.sh
-
Start the training thread by running: sh train.sh
-
Start the segmentation thread by running: sh segment.sh
-
Access the UI by launching the following URL on a browser: http://localhost:8888/browse
Then select a project from the drop down list. Press the start button to activate a project or stop to deactivate. Only one project can be active at a time.