We propose a multi-layer variational autoencoder method, we call HR-VQVAE, that learns hierarchical discrete representations of the data. By utilizing a novel objective function, each layer in HR-VQVAE learns a discrete representation of the residual from previous layers through a vector quantized encoder. Furthermore, the representations at each layer are hierarchically linked to those at previous layers. We evaluate our method on the tasks of image reconstruction and generation. Experimental results demonstrate that the discrete representations learned by HR-VQVAE enable the decoder to reconstruct high-quality images with less distortion than the baseline methods, namely VQVAE and VQVAE-2. HR-VQVAE can also generate high-quality and diverse images that outperform state-of-the-art generative models, providing further verification of the efficiency of the learned representations. The hierarchical nature of HR-VQVAE i) reduces the decoding search time, making the method particularly suitable for high-load tasks and ii) allows to increase the codebook size without incurring the codebook collapse problem.

Adiban M., Stefanov K., Siniscalchi S.M., Salvi G. (2022). Hierarchical Residual Learning Based Vector Quantized Variational Autoencoder for Image Reconstruction and Generation. In BMVC 2022 - 33rd British Machine Vision Conference Proceedings. British Machine Vision Association, BMVA.

Hierarchical Residual Learning Based Vector Quantized Variational Autoencoder for Image Reconstruction and Generation

Siniscalchi S. M.
Supervision
;
2022-01-01

Abstract

We propose a multi-layer variational autoencoder method, we call HR-VQVAE, that learns hierarchical discrete representations of the data. By utilizing a novel objective function, each layer in HR-VQVAE learns a discrete representation of the residual from previous layers through a vector quantized encoder. Furthermore, the representations at each layer are hierarchically linked to those at previous layers. We evaluate our method on the tasks of image reconstruction and generation. Experimental results demonstrate that the discrete representations learned by HR-VQVAE enable the decoder to reconstruct high-quality images with less distortion than the baseline methods, namely VQVAE and VQVAE-2. HR-VQVAE can also generate high-quality and diverse images that outperform state-of-the-art generative models, providing further verification of the efficiency of the learned representations. The hierarchical nature of HR-VQVAE i) reduces the decoding search time, making the method particularly suitable for high-load tasks and ii) allows to increase the codebook size without incurring the codebook collapse problem.
2022
Settore ING-INF/05 - Sistemi Di Elaborazione Delle Informazioni
Adiban M., Stefanov K., Siniscalchi S.M., Salvi G. (2022). Hierarchical Residual Learning Based Vector Quantized Variational Autoencoder for Image Reconstruction and Generation. In BMVC 2022 - 33rd British Machine Vision Conference Proceedings. British Machine Vision Association, BMVA.
File in questo prodotto:
File Dimensione Formato  
Hierarchical+Residual+Learning+Based+Vector+Quantized+Variational+Autoencoder+for+Image+Reconstruction+and+Generation-1-compresso.pdf

accesso aperto

Tipologia: Pre-print
Dimensione 363.39 kB
Formato Adobe PDF
363.39 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10447/637533
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 2
  • ???jsp.display-item.citation.isi??? ND
social impact