Research

Reconstruction

Introduction

Medical image 3D reconstruction is a crucial step in constructing Virtual Physiological Human (VPH), particularly in establishing biomechanical models of the brain. This process involves utilizing Magnetic Resonance (MR) imaging data to measure the 3D geometric structures of brain parenchyma and ventricular systems, subsequently reconstructing a mesh model of the brain. The reconstruction of medical images in 3D can be achieved through open-source software, involving image segmentation and reconstruction output.
During the segmentation phase, a semi-automatic algorithm is employed to acquire velocity images. Image classification algorithms such as Gaussian Mixture Model-based clustering, supervised Random Forest classification, active contour models, and edge detection are utilized. This segmentation categorizes the images into foreground and background, where the foreground represents the region of interest with a value of 1 in velocity images, while the background, representing uninterested areas, holds a value of 0.
In the output reconstruction phase, ensuring voxel dimensions approximate reality is crucial. Therefore, based on prior labeling results, open-source software generates surface meshes. The illustration below depicts 3D human and rat brains successfully established through this reconstruction process.