Home >

news Help

Publication Information


Title
Japanese: 
English:Learning VAE with Categorical Labels for Generating Conditional Handwritten Characters 
Author
Japanese: Goto Keita, 井上 中順.  
English: Keita Goto, Nakamasa Inoue.  
Language English 
Journal/Book name
Japanese: 
English: 
Volume, Number, Page        
Published date July 27, 2021 
Publisher
Japanese: 
English: 
Conference name
Japanese: 
English:17th International Conference on Machine Vision Applications(MVA) 
Conference site
Japanese: 
English: 
Official URL http://www.mva-org.jp/mva2021/
 
Abstract The variational autoencoder (VAE) has succeeded in learning disentangled latent representations from data without supervision. Well disentangled representations can express interpretable semantic value, which is useful for various tasks, including image generation. However, the conventional VAE model is not suitable for data generation with specific category labels because it is challenging to ac- quire categorical information as latent variables. There- fore, we propose a framework for learning label represen- tations in a VAE by using supervised categorical labels as- sociated with data. Through experiments, we show that this framework is useful for generating data belonging to a spe- cific category. Furthermore, we found that our framework successfully disentangled latent factors from similar data of different classes.

©2007 Tokyo Institute of Technology All rights reserved.