| Ratings/Long-term Prospects Comments: | Overall -- Outstanding, fully mature tool. It can easily compete with the best commercial tools. The use is very intuitive, with excellent online help, and easy to learn. Installation -- No problem on all platforms, I tested them all. Documentation -- Outstanding, very detailed online help and full manual. In addition, the NeuroImage paper provides a lot of details about the science and validation. Overall, the documentation might serve as an excellent example how other software should be documented. Long-term -- This is far the best tool for any interactive, user-guided segmentation of voumetric data. The 3D snake is the best example how a well-known computer vision tool can be integrated into a GUI so that it can be distributed to the broad community. This is a tool for everyone who needs to quickly, efficiently and reliably segment structures in 3D. It is thus useful for a very broad category and not limited to MRI, fMRI etc., but to any application with volumetric data.
A possible extension could be application to much larger datasets, where users can select a subvolume and then apply the processing to this region. This processing would finally be integrated back into the much bigger volume. This might have applications in microscopy imaging, e.g. Feature Requests -- I would urge the developers not to add too much. The tool has a limited scope and thus do the job it is designed for extremely well, and training usually takes only 30'. Other tools try to solve everything and are thus very difficult to learn and to maintain. |
