search

microsoft / Recursive-Cascaded-Networks

[ICCV 2019] Recursive Cascaded Networks for Unsupervised Medical Image Registration
https://arxiv.org/abs/1907.12353
MIT License
361 stars 87 forks source link

In the liver.py, i see the following code. #40

Closed krisice closed 3 years ago

krisice commented 3 years ago 
        while True:
            ret = dict()
            ret['voxel1'] = np.zeros(
                (batch_size, 128, 128, 128, 1), dtype=np.float32)
            ret['voxel2'] = np.zeros(
                (batch_size, 128, 128, 128, 1), dtype=np.float32)
            ret['seg1'] = np.zeros(
                (batch_size, 128, 128, 128, 1), dtype=np.float32)
            ret['seg2'] = np.zeros(
                (batch_size, 128, 128, 128, 1), dtype=np.float32)
            ret['point1'] = np.ones(
                (batch_size, np.sum(valid_mask), 3), dtype=np.float32) * (-1)
            ret['point2'] = np.ones(
                (batch_size, np.sum(valid_mask), 3), dtype=np.float32) * (-1)
            ret['id1'] = np.empty((batch_size), dtype='<U40')
            ret['id2'] = np.empty((batch_size), dtype='<U40')

The ret['voxle']、ret['seg'] and ['point'] what does they mean? The train data is v[voxel]?