You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Practically in our experiments, we empirically adopt the window size as 10, the sliding stride as 2, and the number of bands as M = 6, thus the size of the output matrix will be 149 × 149 × 6 if the input image is of size 299 × 299 × 3.
Practically in our experiments, we empirically adopt the window size as 10, the sliding stride as 2, and the number of bands as M = 6, thus the size of the output matrix will be 149 × 149 × 6 if the input image is of size 299 × 299 × 3.
我比较疑惑的是,在FAD中的输入是3通道图做DCT,
x_freq = self._DCT_all @ x @ self._DCT_all_T # [N, 3, 299, 299]
但是在LFS时,却要先转成灰度图:
x_gray = 0.299x[:,0,:,:] + 0.587x[:,1,:,:] + 0.114*x[:,2,:,:]
x = x_gray.unsqueeze(1)
x = (x + 1.) * 122.5
文章中,两块应该都是RGB输入,请问为什么LFS时要灰度图输入啊??
The text was updated successfully, but these errors were encountered: