Rethinking the Detail-Preserved Completion of Complex Tubular Structures based on Point Cloud: a Dataset and a Benchmark

TL;DR

This paper introduces a new dataset and benchmark for tubular structure completion in medical imaging, proposing a novel neural network that improves the accuracy and integrity of reconnection in complex point cloud data.

Contribution

The study pioneers tubular structure completion using point cloud data, establishing a new benchmark and proposing TSRNet, a network that enhances structural reconnection in medical imaging.

Findings

TSRNet outperforms existing methods on multiple datasets

The PC-CAC dataset provides a valuable benchmark for future research

The proposed method maintains structural integrity during reconstruction

Abstract

Complex tubular structures are essential in medical imaging and computer-assisted diagnosis, where their integrity enhances anatomical visualization and lesion detection. However, existing segmentation algorithms struggle with structural discontinuities, particularly in severe clinical cases such as coronary artery stenosis and vessel occlusions, which leads to undesired discontinuity and compromising downstream diagnostic accuracy. Therefore, it is imperative to reconnect discontinuous structures to ensure their completeness. In this study, we explore the tubular structure completion based on point cloud for the first time and establish a Point Cloud-based Coronary Artery Completion (PC-CAC) dataset, which is derived from real clinical data. This dataset provides a novel benchmark for tubular structure completion. Additionally, we propose TSRNet, a Tubular Structure Reconnection Network that integrates a detail-preservated feature extractor, a multiple dense refinement strategy, and a global-to-local loss function to ensure accurate reconnection while maintaining structural integrity. Comprehensive experiments on our PC-CAC and two additional public datasets (PC-ImageCAS and PC-PTR) demonstrate that our method consistently outperforms state-of-the-art approaches across multiple evaluation metrics, setting a new benchmark for point cloud-based tubular structure reconstruction. Our benchmark is available at https://github.com/YaoleiQi/PCCAC.