Capturing implicit hierarchical structure in 3D biomedical images with self-supervised hyperbolic representations

Joy Hsu, Jeffrey Gu, Gong Her Wu, Wah Chiu, Serena Yeung

Keywords: hyperbolic, hierarchy, representation learning, unsupervised segmentation, computer vision, biomedical

TL;DR: Leveraging hyperbolic representations to capture hierarchical structure in biomedical images for segmentation.

Abstract: We consider the task of representation learning for unsupervised segmentation of 3D voxel-grid biomedical images. We show that models that capture implicit hierarchical relationships between subvolumes are better suited for this task. To that end, we consider encoder-decoder architectures with a hyperbolic latent space, to explicitly capture hierarchical relationships present in subvolumes of the data. We propose utilizing a 3D hyperbolic variational autoencoder with a novel gyroplane convolutional layer to map from the embedding space back to 3D images. To capture these relationships, we introduce an essential self-supervised loss---in addition to the standard VAE loss---which infers approximate hierarchies and encourages implicitly related subvolumes to be mapped closer in the embedding space. We present experiments on synthetic datasets along with a dataset from the medical domain to validate our hypothesis.

Code Of Conduct: I certify that all co-authors of this work have read and commit to adhering to the NeurIPS Statement on Ethics, Fairness, Inclusivity, and Code of Conduct.