Style Transfer and Extraction for the Handwritten Letters Using Deep Learning

Abstract : How can we learn, transfer and extract handwriting styles using deep neural networks? This paper explores these questions using a deep conditioned autoencoder on the IRON-OFF handwriting data-set. We perform three experiments that systematically explore the quality of our style extraction procedure. First, We compare our model to handwriting benchmarks using multidimensional performance metrics. Second, we explore the quality of style transfer, i.e. how the model performs on new, unseen writers. In both experiments, we improve the metrics of state of the art methods by a large margin. Lastly, we analyze the latent space of our model, and we see that it separates consistently writing styles.
Complete list of metadatas

Cited literature [20 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02049006
Contributor : Gérard Bailly <>
Submitted on : Tuesday, February 26, 2019 - 9:10:44 AM
Last modification on : Thursday, April 11, 2019 - 4:18:05 PM
Long-term archiving on : Monday, May 27, 2019 - 12:51:10 PM

File

osm_ICAART2019_org.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-02049006, version 1

Citation

Omar Mohammed, Gérard Bailly, Damien Pellier. Style Transfer and Extraction for the Handwritten Letters Using Deep Learning. 11th International Conference on Agents and Artificial Intelligence (ICAART 2019), Feb 2019, Prague, Czech Republic. ⟨hal-02049006⟩

Share

Metrics

Record views

35

Files downloads

94