 |
 |
Hi, The model that we distribute is very different from the training set, which consists of the original images and their manual segmentation. By comparison, the model just represents the mean and eigenvectors of the shape and appearance in the vicinity of the subcortical structures. We can easily distribute the models, but not the images in the training set. Also, I would recommend using the models in FSL4.1 and not the old ones. All the best, Mark On 23 Jan 2009, at 19:22, Hedok Lee wrote: > Dear Mark and Siamak, > > I was wondering about the same thing as the latest verison of FSL > does not include model 317. But I happened to keep a copy of both > FSL 4.0 and FSL 4.1, so what if we just copy the files from 4.0 to > 4.1? Will that work? > Siamak, > I can send you the model from 317 (FSL 4.0) if you want to try. > > > Hedok > > > Mark Jenkinson wrote: >> Hi, >> >> No - the dataset is not available for downloading. >> A small subset of it overlaps with the IBSR data >> which is publically available at: >> http://www.cma.mgh.harvard.edu/ibsr/data.html >> >> All the best, >> Mark >> >> >> On 23 Jan 2009, at 17:29, siamak wrote: >> >>> Hi, >>> >>> I'd like to know that the training data (317 training subject and >>> their manually >>> segmented subcortical structures) used to implement the FIRST tool >>> is >>> available for downloading or not. >>> >>> Siamak >>> >