Does bicubic downsampling in PyTorch torch.nn.functional.interpolate internally involve anti-aliasing? If not, how can we avoid aliasing when direct using bicubic downsample?

In Matlab, I know one can use imresize to downscale an image (use bicubic as default) and setting 'anitaliasing=true' to avoid aliasing. However, there is no similar option in torch.functional.interpolate(mode='bicubic'). I can't find out more details from the source code. Does anybody know that? Many thanks.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.