Saved in:
Bibliographic Details
Main Authors: Sundgaard, Josefine Vilsbøll, Hannemose, Morten Rieger, Laugesen, Søren, Bray, Peter, Harte, James, Kamide, Yosuke, Tanaka, Chiemi, Paulsen, Rasmus R., Christensen, Anders Nymark
Format: Preprint
Published: 2022
Subjects:
Online Access:https://arxiv.org/abs/2202.03434
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913743789621248
author Sundgaard, Josefine Vilsbøll
Hannemose, Morten Rieger
Laugesen, Søren
Bray, Peter
Harte, James
Kamide, Yosuke
Tanaka, Chiemi
Paulsen, Rasmus R.
Christensen, Anders Nymark
author_facet Sundgaard, Josefine Vilsbøll
Hannemose, Morten Rieger
Laugesen, Søren
Bray, Peter
Harte, James
Kamide, Yosuke
Tanaka, Chiemi
Paulsen, Rasmus R.
Christensen, Anders Nymark
contents We present a deep metric variational autoencoder for multi-modal data generation. The variational autoencoder employs triplet loss in the latent space, which allows for conditional data generation by sampling in the latent space within each class cluster. The approach is evaluated on a multi-modal dataset consisting of otoscopy images of the tympanic membrane with corresponding wideband tympanometry measurements. The modalities in this dataset are correlated, as they represent different aspects of the state of the middle ear, but they do not present a direct pixel-to-pixel correlation. The approach shows promising results for the conditional generation of pairs of images and tympanograms, and will allow for efficient data augmentation of data from multi-modal sources.
format Preprint
id arxiv_https___arxiv_org_abs_2202_03434
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Multi-modal data generation with a deep metric variational autoencoder
Sundgaard, Josefine Vilsbøll
Hannemose, Morten Rieger
Laugesen, Søren
Bray, Peter
Harte, James
Kamide, Yosuke
Tanaka, Chiemi
Paulsen, Rasmus R.
Christensen, Anders Nymark
Image and Video Processing
Computer Vision and Pattern Recognition
We present a deep metric variational autoencoder for multi-modal data generation. The variational autoencoder employs triplet loss in the latent space, which allows for conditional data generation by sampling in the latent space within each class cluster. The approach is evaluated on a multi-modal dataset consisting of otoscopy images of the tympanic membrane with corresponding wideband tympanometry measurements. The modalities in this dataset are correlated, as they represent different aspects of the state of the middle ear, but they do not present a direct pixel-to-pixel correlation. The approach shows promising results for the conditional generation of pairs of images and tympanograms, and will allow for efficient data augmentation of data from multi-modal sources.
title Multi-modal data generation with a deep metric variational autoencoder
topic Image and Video Processing
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2202.03434