HIGHLIGHTS
- who: Andreas M. Kist from the Department Artificial Intelligence in Biomedical Engineering, Friedrich-Alexander-University Erlangen-Nürnberg have published the Article: A single latent channel is sufficient for biomedical glottis segmentation, in the Journal: Scientific Reports Scientific Reports
- what: The authors use a reference implementation with 16 filters in the first layer ( fbase=16) and thus, 256 channels in the latent space, as this has been shown previously to provide comparable performance compared to fL=102410. In Supplementary Figure 1, the authors show that the autoencoder ability across latent space channels are perceptually uniform . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.