Sparse Representation Learning for Vessels

2026-05-02

AI Summary

Existing methods struggle to efficiently model clinical-grade whole-organ vascular networks, often constrained by limited spatial coverage or structural simplifications. This work proposes VAEsselSparse, a novel model that achieves the first efficient sparse representation of whole-organ vasculature at sub-millimeter resolution. Built upon a variational autoencoder framework, VAEsselSparse integrates sparse convolutions and attention mechanisms to exploit the inherent sparsity of vascular structures, enabling an 8×8×8 spatial compression. The method surpasses both dense models and current approaches in reconstruction fidelity, while its compact latent space effectively supports downstream clinical tasks such as aneurysm and stenosis classification. Furthermore, the learned representation serves as a generative prior for synthesizing anatomically realistic vascular structures, successfully balancing high-fidelity reconstruction with practical clinical applicability.

Abstract

Analyzing human vasculature and vessel-like, tubular structures, such as airways, is crucial for disease diagnosis and treatment. Current methods often rely on small sub-regions or simplified tree-like structures, rendering analysis of entire organ-level networks at clinical resolution computationally challenging. To this end, we propose VAEsselSparse, an efficient encoder-decoder model to obtain a meaningful yet compact representation of the entire organ-level vascular network at sub-millimeter resolution. VAEsselSparse leverages the inherent sparsity of 3D vascular structures via sparse convolutions and attention mechanisms, achieving substantial spatial compression rates of 8 x 8 x 8. We demonstrate superior reconstruction performance compared to dense counterparts and previous methods. Importantly, the resulting latent space retains clinically relevant discriminative features readily usable for classification tasks, such as aneurysm/stenosis or subvariants of the circle of Willis. Moreover, the compact latent space of VAEsselSparse serves as an effective representation for learning vessel-specific priors through generative models, enabling the synthesis of realistic vasculature.

Problem

• Sparse Representation
• Vascular Networks
• Organ-level Analysis
• 3D Vasculature
• Clinical Resolution

Innovation

• sparse convolution
• attention mechanism
• vascular representation learning
• latent space compression
• generative modeling