A latent space exploration for microscopic skin lesion augmentations with VQ-VAE-2 and PixelSNAIL

Alessio Gallucci, Nicola Pezzotti, Dmitry Znamenskiy, Milan Petkovic

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review


Skin cancer affects more than 3 million people only in the US. Comprehensive microscopic databases include around 30 thousand samples, limiting the richness of patterns that can be presented to machine learning. To this end, generative models such as GANs have been proposed for creating realistic synthetic images but, despite their popularity, they are often difficult to train and control. Recently an autoregressive approach based on a quantized autoencoder showed state of the art performances while being simple to train and provide synthetic data generation opportunities. In the first part of this paper we evaluate the training of VQ-VAE-2 with different latent space configuration. In the second part, we show how to use a learned prior over the latent space with PixelSNAIL to generate and modify skin lesions. We show how this process can be used for powerful data augmentation and visualization for skin health, evaluating it on a downstream application that classifies malignant lesions

Original languageEnglish
Title of host publicationMedical Imaging 2021
Subtitle of host publicationImage Processing
EditorsIvana Isgum, Bennett A. Landman
ISBN (Electronic)9781510640214
Publication statusPublished - 2021
EventMedical Imaging 2021: Image Processing - Virtual, Online, United States
Duration: 15 Feb 202119 Feb 2021

Publication series

NameProceedings of SPIE
ISSN (Print)1605-7422


ConferenceMedical Imaging 2021: Image Processing
Country/TerritoryUnited States
CityVirtual, Online

Bibliographical note

Publisher Copyright:
© 2021 SPIE.


  • Autoencoders
  • Autoregressive generation
  • Dermatology
  • Image generation
  • Skin lesions


Dive into the research topics of 'A latent space exploration for microscopic skin lesion augmentations with VQ-VAE-2 and PixelSNAIL'. Together they form a unique fingerprint.

Cite this