Deformable convolution in 2d and 3d
Usage
from deformable_conv_3d import deform_conv3d
image = torch.rand(1, 16, 64, 64, 64)
def_conv3 = deform_conv3d(in_c=16, out_c=32, kernel_size=(3,3,3))
result = def_conv3(image)
Pytorch implementation of 2D and 3D deformable convolution specified in https://arxiv.org/abs/1703.06211.
from deformable_conv_3d import deform_conv3d
image = torch.rand(1, 16, 64, 64, 64)
def_conv3 = deform_conv3d(in_c=16, out_c=32, kernel_size=(3,3,3))
result = def_conv3(image)
Pytorch implementation of 2D and 3D deformable convolution specified in https://arxiv.org/abs/1703.06211.
MIT License