| Description: | Cardiac biventricular mesh models |
| Category: | Assessment centre ⏵ Imaging ⏵ Heart MRI ⏵ Cardiac 3D meshes |
| Participants | 53,224 | Value Type | Text | Sexed | Both sexes | Debut | Aug 2025 | |||
| Item count | 53,224 | Item Type | Bulk | Instances | Defined (1) | Version | Aug 2025 | |||
| Stability | Complete | Strata | Derived | Array | No | Cost Tier | o2 s3 |
WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
The values held within this field are coded links to files held within a bulk repository, not the actual data itself.
The surface meshes are constructed using the methods described in Ugurlu et al. (2025). They are derived from automatic segmentation of the short- and long- axis cine heart MRI scans. There are typically 50 models per case, one per frame. Some frames may be missing, due to failure of the segmentation process to produce sufficient 3D contour and landmark information to construct the mesh. Due to lack of RV myocardium segmentation labels, the RV epicardial surface is generated by dilating the endocardial surface by a fixed thickness.
The number of frames available can be found in Field 30119.
Quality control test results are described in Ugurlu et al. (2025).
The biventricular surface meshes are similar to those described in Publication 2843 and Burns et al. (2025). Each model comprises 388 control points in DICOM Patient coordinates (x,y,z coordinates in mm). This "coarse mesh" must be subdivided to visualise the final surface. Python code scripts for converting the control mesh points into .obj, and other formats, including volume meshes, can be found here.
Enabling scientific discoveries that improve human health