The establishment of static digital humans and the integration with spinal models

TL;DR

AIS presents a 3D spinal deformity whose dynamic behavior cannot be captured by static imaging. The authors propose a pipeline that builds a static digital human representation by fusing multi-view images with CT-derived spine data, using a 3D Gaussian point cloud integrated with SMPL and a genus-0 skin surface, then aligning to a standard skeletal model via feature-point registration and ARAP refinement. The work delivers a multimodal AIS dataset, a precise static spine-integrated representation, and Cobb-angle validation showing sub-degree accuracy against upright X-ray measurements, providing a solid foundation for future dynamic digital human simulations, clinical diagnosis, and surgical planning. This framework enables accurate, patient-specific anatomy visualization and sets the stage for dynamic spine analyses under realistic body postures and movements.

Abstract

Adolescent idiopathic scoliosis (AIS), a prevalent spinal deformity, significantly affects individuals' health and quality of life. Conventional imaging techniques, such as X - rays, computed tomography (CT), and magnetic resonance imaging (MRI), offer static views of the spine. However, they are restricted in capturing the dynamic changes of the spine and its interactions with overall body motion. Therefore, developing new techniques to address these limitations has become extremely important. Dynamic digital human modeling represents a major breakthrough in digital medicine. It enables a three - dimensional (3D) view of the spine as it changes during daily activities, assisting clinicians in detecting deformities that might be missed in static imaging. Although dynamic modeling holds great potential, constructing an accurate static digital human model is a crucial initial step for high - precision simulations. In this study, our focus is on constructing an accurate static digital human model integrating the spine, which is vital for subsequent dynamic digital human research on AIS. First, we generate human point - cloud data by combining the 3D Gaussian method with the Skinned Multi - Person Linear (SMPL) model from the patient's multi - view images. Then, we fit a standard skeletal model to the generated human model. Next, we align the real spine model reconstructed from CT images with the standard skeletal model. We validated the resulting personalized spine model using X - ray data from six AIS patients, with Cobb angles (used to measure the severity of scoliosis) as evaluation metrics. The results indicate that the model's error was within 1 degree of the actual measurements. This study presents an important method for constructing digital humans.

Figures (9)

• Figure 1: The framework of the proposed method.
• Figure 2: In the given frame, sampling operations are carried out on the human body surface to form a point cloud, and at the same time, the corresponding positions of these points are marked in the UV position map. After that, the UV position map is fed into the pose encoder to generate the corresponding pose features. During this period, the optimizable feature tensor is precisely aligned with the pose features in a specific way, with the aim of more effectively capturing the overall appearance of the human body. These aligned features are input into the Gaussian parameter decoder, which can predict the offset $\delta x$, color c and scale s of each point. And these predicted results, together with the fixed rotation q and opacity $\alpha$, jointly form an animatable 3D Gaussian distribution in the canonical space.
• Figure 3: SMPL 3D model and skeletal landmarks
• Figure 4: Skin surface $M$
• Figure 5: A fitted and standard skeleton model based on the SMPL 3D model