A Strong Linear Baseline for Whole-Heart Cardiac Shape Completion on CT, with an Open Eleven-Structure Statistical Shape Model
Researchers have developed a statistical shape model for the heart that can be used to complete missing structures in cardiac computed-tomography (CT) scans. The model is built from 383 automatically labelled cases and uses a closed-form conditional-Gaussian estimator to reconstruct missing parts. In experiments, this method performed better than other approaches on several benchmarks, including a second public dataset. However, the authors note that their model is not suitab
Researchers have developed a statistical shape model for the heart that can be used to complete missing structures in cardiac computed-tomography (CT) scans. The model is built from 383 automatically labelled cases and uses a closed-form conditional-Gaussian estimator to reconstruct missing parts. In experiments, this method performed better than other approaches on several benchmarks, including a second public dataset. However, the authors note that their model is not suitable for clinical use but rather for cohort-unification research.
---
Why it matters: This work matters because it provides a new tool for researchers working with cardiac CT scans, allowing them to pool data from different sources and study the heart's shape in more detail. This could lead to better understanding of cardiovascular diseases and improved diagnosis methods.
Source: https://arxiv.org/abs/2608.19932
This article was originally published at: https://arxiv.org/abs/2608.19932