Papers with Japanese character images
Handwritten Character Generation using Y-Autoencoder for Character Recognition Model Training (2022.lrec-1)
Copied to clipboard
| Challenge: | re-emergence of deep learning since third winter of artificial intelligence has led to mainstreaming of deep-learning systems that use large amounts of data to train a model. |
| Approach: | They propose a Y-Autoencoder-based handwritten character generator to generate Japanese Hiragana characters with a single image to increase the amount of data needed for character recognition. |
| Outcome: | The proposed system generates Japanese Hiragana characters with a single image . the results show that the Y-AE-based generator produces an improved F1 score . |