Skip to content

Latest commit

 

History

History
48 lines (26 loc) · 1.19 KB

README.md

File metadata and controls

48 lines (26 loc) · 1.19 KB

Description

This code repository is the official implementation of the paper "DCC: Dynamic Category Compression for Enhanced Handwritten Text Generation".

Instruction

Our proposed method is generic and effective, and theoretically can be seamlessly integrated into existing HTG methods.

For the relevant code, please check the network section of the corresponding method on your own.

Generalized HTG method based on GAN

Model

Our method

Model

Result

Result

Pre-trained models

The latest SOTA model will be released after the paper is received.

Evaluation metrics

FID：

https://github.com/mseitzer/pytorch-fid

HWD, KID：

https://github.com/aimagelab/HWD

Related Code

https://github.com/aimagelab/VATr

https://github.com/ankanbhunia/Handwriting-Transformers

https://github.com/omni-us/research-GANwriting

https://github.com/EDM-Research/VATr-pp

https://github.com/lijian16/FCC

Acknowledgements

Thanks to Kang Lei, Ankan Kumar Bhunia, and Vittorio Pippi for their contributions to the HTG field.