Abstract:Scoliosis is a disease caused by deformation of the spine, and precise imaging of human spine tissues is crucial and instructive for the subsequent treatment. Currently, radiographs are clinically used to monitor disease progress. However, radiation exposure is a safety concern and is undesirable for regular monitoring during rehabilitation intervention. In this study, we developed a novel technique to generate the 3D structure of human spine from a tracked freehand ultrasound scanning. First of all, we designed an approach for detection and location of every vertebra in the spine by firstly detecting the vertebral landmarks from the ultrasound B-scan sequence, then computing the size, location and posture of every vertebra and finally using a method of interpolation to form the whole spine according to the prior knowledge of vertebral anatomical structure. Deep learning based object detection methods were used to find the landmarks of the vertebrae, and those landmarks were then clustered to model every vertebra (i.e. computing its location, posture and real size). Due to the situations where the vertebra is not clearly seen, we therefore used the technique of interpolation to estimate the location of those missing vertebrae with respect to the anatomical prior. In addition, we reconnected all vertebral models and reappeared the whole spine in 3D space. With that, the Cobb angle could be easily measured. The phantom and in vivo experiments were also conducted to further verify the precision of the proposed approaches. The results have demonstrated its feasibility in the clinical practice of adolescent idiopathic scoliosis patients (R2 = 0.9780,p < 0.001), which is beneficial for the regular quantitative monitoring of disease progression.

A Novel Skeleton-based Model with Spine for 3D Human Pose Estimation

Animating Scanned Human Models.

Real-time Skeletonization for Sketch-Based Modeling

Mmskeleton: 3D Human Skeleton Estimation Using Millimeter Wave Radar Sparse Point Clouds

Symmetry-aware Kinematic Skeleton Generation of a 3D Human Body Model.

Human Pose Tracking Algorithm Based on Skeleton-Texture Model

Reconstruction of a Human Spine Curve in the Sagittal Plane From the Measured Contour of the Human Back

Physically Based Real-Time Human Figure Animation

3D Spine Model Reconstruction Based on RGBD Images of Unclothed Back Surface

A 3D 2-subiteration thinning algorithm for human pose estimation

3D Articulated Skeleton Extraction Using a Single Consumer-Grade Depth Camera.

Dynamic Human Body Reconstruction and Motion Tracking with Low-Cost Depth Cameras

3D Kinematics Estimation from Video with a Biomechanical Model and Synthetic Training Data

Estimation of Human Body Shape and Posture Under Clothing

A method of sketch-based human body modeling

Anatomically Detailed Simulation of Human Torso

Anatomical prior based vertebra modelling for reappearance of human spines

Predicting Animation Skeletons for 3D Articulated Models via Volumetric Nets

DeepSkeleton: Skeleton Map for 3D Human Pose Regression.

Unsupervised 3D Pose Estimation with Non-Rigid Structure-from-Motion Modeling

A generic detailed multibody model for simulating thoracic spine and ribcage kinematics