Hypercomplex generative adversarial networks for lightweight semantic labeling
Following recent advances on parameterized hypercomplex multiplication, we explore the
usefulness of hypercomplex convolutions and deconvolutions in a document labeling task …
usefulness of hypercomplex convolutions and deconvolutions in a document labeling task …
Keyword spotting with quaternionic ResNet: application to spotting in Greek manuscripts
Quaternionized versions of standard (real-valued) neural network layers have shown to lead
to networks that are sparse and as effective as their real-valued counterparts. In this work …
to networks that are sparse and as effective as their real-valued counterparts. In this work …
On-the-fly deformations for keyword spotting
Abstract Modern Keyword Spotting systems rely on deep learning approaches to build
effective neural networks which provide state-of-the-art results. Despite their evident …
effective neural networks which provide state-of-the-art results. Despite their evident …
[PDF][PDF] Exploring Uses of Normalizing Flows for Document Image Processing: Text Super-Resolution and Binarization
Exploring Uses of Normalizing Flows for Document Image Processing: Text Super-Resolution
and Binarization Page 1 Exploring Uses of Normalizing Flows for Document Image Processing …
and Binarization Page 1 Exploring Uses of Normalizing Flows for Document Image Processing …