SepVAE: a contrastive VAE to separate pathological patterns from healthy ones

Robin Louiset, Edouard Duchesnay, Grigis Antoine, Benoit Dufumier, Pietro Gori
Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning, PMLR 250:918-936, 2024.

Abstract

Contrastive Analysis VAE (CA-VAEs) is a family of Variational auto-encoders (VAEs) that aims at separating the common factors of variation between a \textit{background} dataset (BG) (\textit{i.e.,} healthy subjects) and a \textit{target} dataset (TG) (\textit{i.e.,} patients) from the ones that only exist in the target dataset. To do so, these methods separate the latent space into a set of \textbf{salient} features (\textit{i.e.,} proper to the target dataset) and a set of \textbf{common} features (\textit{i.e.,} exist in both datasets). Currently, all CA-VAEs models fail to prevent sharing of information between the latent spaces and to capture all salient factors of variation. To this end, we introduce two crucial regularization losses: a disentangling term between common and salient representations and a classification term between background and target samples in the salient space. We show a better performance than previous CA-VAEs methods on three medical applications and a natural images dataset (CelebA).

Cite this Paper

BibTeX
@InProceedings{pmlr-v250-louiset24a, title = {SepVAE: a contrastive VAE to separate pathological patterns from healthy ones}, author = {Louiset, Robin and Duchesnay, Edouard and Antoine, Grigis and Dufumier, Benoit and Gori, Pietro}, booktitle = {Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning}, pages = {918--936}, year = {2024}, editor = {Burgos, Ninon and Petitjean, Caroline and Vakalopoulou, Maria and Christodoulidis, Stergios and Coupe, Pierrick and Delingette, Hervé and Lartizien, Carole and Mateus, Diana}, volume = {250}, series = {Proceedings of Machine Learning Research}, month = {03--05 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v250/main/assets/louiset24a/louiset24a.pdf}, url = {https://proceedings.mlr.press/v250/louiset24a.html}, abstract = {Contrastive Analysis VAE (CA-VAEs) is a family of Variational auto-encoders (VAEs) that aims at separating the common factors of variation between a \textit{background} dataset (BG) (\textit{i.e.,} healthy subjects) and a \textit{target} dataset (TG) (\textit{i.e.,} patients) from the ones that only exist in the target dataset. To do so, these methods separate the latent space into a set of \textbf{salient} features (\textit{i.e.,} proper to the target dataset) and a set of \textbf{common} features (\textit{i.e.,} exist in both datasets). Currently, all CA-VAEs models fail to prevent sharing of information between the latent spaces and to capture all salient factors of variation. To this end, we introduce two crucial regularization losses: a disentangling term between common and salient representations and a classification term between background and target samples in the salient space. We show a better performance than previous CA-VAEs methods on three medical applications and a natural images dataset (CelebA).} }
Endnote
%0 Conference Paper %T SepVAE: a contrastive VAE to separate pathological patterns from healthy ones %A Robin Louiset %A Edouard Duchesnay %A Grigis Antoine %A Benoit Dufumier %A Pietro Gori %B Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2024 %E Ninon Burgos %E Caroline Petitjean %E Maria Vakalopoulou %E Stergios Christodoulidis %E Pierrick Coupe %E Hervé Delingette %E Carole Lartizien %E Diana Mateus %F pmlr-v250-louiset24a %I PMLR %P 918--936 %U https://proceedings.mlr.press/v250/louiset24a.html %V 250 %X Contrastive Analysis VAE (CA-VAEs) is a family of Variational auto-encoders (VAEs) that aims at separating the common factors of variation between a \textit{background} dataset (BG) (\textit{i.e.,} healthy subjects) and a \textit{target} dataset (TG) (\textit{i.e.,} patients) from the ones that only exist in the target dataset. To do so, these methods separate the latent space into a set of \textbf{salient} features (\textit{i.e.,} proper to the target dataset) and a set of \textbf{common} features (\textit{i.e.,} exist in both datasets). Currently, all CA-VAEs models fail to prevent sharing of information between the latent spaces and to capture all salient factors of variation. To this end, we introduce two crucial regularization losses: a disentangling term between common and salient representations and a classification term between background and target samples in the salient space. We show a better performance than previous CA-VAEs methods on three medical applications and a natural images dataset (CelebA).
APA
Louiset, R., Duchesnay, E., Antoine, G., Dufumier, B. & Gori, P.. (2024). SepVAE: a contrastive VAE to separate pathological patterns from healthy ones. Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 250:918-936 Available from https://proceedings.mlr.press/v250/louiset24a.html.

Related Material