“HiGAN+: Handwriting Imitation GAN With Disentangled Representations” by Gan, Wang, Leng and Gao
Conference:
Type(s):
Title:
- HiGAN+: Handwriting Imitation GAN With Disentangled Representations
Session/Category Title: Magical Sketching
Presenter(s)/Author(s):
Moderator(s):
Abstract:
While this task is challenging for machines, humans can quickly learn handwriting imitation with hallucinations. Humans can imitate handwriting perhaps because they can learn to disentangle textual contents and calligraphic styles. Inspired by this, we propose HiGAN+ for realistic handwritten text synthesis based on disentangled representations, achieving the state-of-the-art performance.