torch.nn.functional.
grid_sample
(input, grid, mode='bilinear', padding_mode='zeros')
我用于image warp的图像恢复
outpuy_{x,y} 的像素值与input_{x0,y0} 的像素值一致, 矩阵索引通过grid矩阵保存。 grid_{x,y}=(x0,y0)
In the spatial (4-D) case, for input
with shape (N,C,Hin,Win)(N, C, H_\text{in}, W_\text{in})(N,C,Hin,Win) and grid
with shape (N,Hout,Wout,2)(N, H_\text{out}, W_\text{out}, 2)(N,Hout,Wout,2), the output will have shape (N,C,Hout,Wout)(N, C, H_\text{out}, W_\text{out})(N,C,Hout,Wout).
input (N,C,H_in,W_in)
grid (N ,H_out,W_out, 2)
grid索引值(数据坐标)对应output索引值(数据坐标)
grid中数值对应input的索引值(数据坐标)
grid(u,v) = x,y
output(u,v)数值对应input(x,y)
示意图如下: