← ArchivePaper2023
BOSS: Bones, Organs and Skin Shape Model
Karthik Shetty, Annette Birkhold, Srikrishna Jaganathan, Norbert Strobel, Bernhard Egger, Markus Kowarschik, Andreas Maier
Deformable full-body model integrating skin, organs, and bones trained on 300 CT scans using SMPL architecture, achieving 3.6 mm bone and 8.8 mm organ accuracy.
Abstract
A virtual anatomical model of a patient can be a valuable tool for enhancing clinical tasks such as workflow automation, patient-specific X-ray dose optimization, markerless tracking, positioning, and navigation assistance in image-guided interventions. For these tasks, it is highly desirable that the patient's surface and internal organs are of high quality for any pose and shape estimate. At present, the majority of statistical shape models (SSMs) are restricted to a small number of organs or bones or do not adequately represent the general population. To address this, we propose a deformable human shape and pose model that combines skin, internal organs, and bones, learned from CT images. By modeling the statistical variations in a pose-normalized space using probabilistic PCA while also preserving joint kinematics, our approach offers a holistic representation of the body that can be beneficial for automation in various medical applications. In an interventional setup, our model could, for example, facilitate automatic system/patient positioning, organ-specific iso-centering, automated collimation or collision prediction. We assessed our model's performance on a registered dataset, utilizing the unified shape space, and noted an average error of 3.6 mm for bones and 8.8 mm for organs.
How to read this
- Category
- Method / model: a full-body statistical shape model with internal anatomy
- Contributions
- A deformable human shape and pose model integrating skin, internal organs, and bones, learned from CT images
- Statistical variation modeled with probabilistic PCA in a pose-normalized space while preserving joint kinematics
- Reported accuracy of 3.6 mm for bone and 8.8 mm for organs, with proposed medical uses such as positioning, iso-centering, and collision prediction
- Context
- Builds on SMPL's skinned shape-model architecture (Loper) and skeletal-from-surface work like OSSO (Keller), extending statistical body models to include organs and bones for clinical use.Builds on: SMPL: A Skinned Multi-Person Linear Model · OSSO: Obtaining Skeletal Shape from Outside
- Correctness
- Trained and assessed on roughly 300 CT scans with stated millimeter accuracy; population coverage and generalization beyond the training cohort, plus pose/shape extremes, are the caveats to keep in mind for the medical claims.
- Clarity
- Accessible if familiar with SMPL-style models; a first pass conveys the integrated-anatomy idea, a second pass covers the probabilistic PCA and registration.
- How to read it
- First pass for what the model integrates and the reported accuracies; do a second pass on the pose-normalized PCA and CT registration if you intend to use or extend it clinically.
Builds on
Built upon by
Nothing yet.
Related work
- From Skin to Skeleton: Towards Biomechanically Accurate 3D Digital Humans 2023 / SIGGRAPH Asia
- NIMBLE: A Non-rigid Hand Model with Bones and Muscles 2022 / TOG
- HIT: Estimating Internal Human Implicit Tissues from the Body Surface 2024 / CVPR
- Animatable Neural Radiance Fields for Modeling Dynamic Human Bodies 2021 / CVPR
Keywords
This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →