Trained FIL model

The FIL model was trained on all 15 training subjects from the MICCAI Challenge dataset, using a minimally informative, but proper Wishart prior, with v0=1.0. An augmentation search radius of 3 voxels was used with a Gaussian weighting standard deviation of 2.0 voxels. Patch sizes were 4x4x4 voxels, and four outer iterations were used for model training. Up to K=24 basis functions were available to encode each patch. This file is the result.

FIL模型在MICCAI挑战赛数据集（MICCAI Challenge Dataset）的全部15个训练受试者上完成训练，采用了最小信息且合规的Wishart先验（Wishart prior），其v0参数取值为1.0。数据增强环节使用了3个体素的搜索半径，并将高斯加权的标准差设置为2.0个体素。训练所用图像块（Patch）尺寸为4×4×4个体素，模型训练共执行4轮外层迭代。每个图像块可调用最多K=24个基函数完成编码。本文件即为该训练流程的输出结果。