Multiscale Modeling and Reconstruction of Joint Motion: Finite Element Optimization Based on Particle Swarm Algorithm

This study employs finite element analysis (FEA) to address the challenge of motion reconstruction in three-dimensional (3D) human body modeling, specifically focusing on joint movement. High-resolution CT and MRI scans of the knee joint were utilized to construct an accurate 3D reconstruction model, while the particle swarm algorithm was implemented to optimize joint positioning. Experimental evaluations were conducted to assess the influence of varying unit numbers and joint rotation angles on reconstruction accuracy. The results demonstrate that increasing the number of units significantly improves the model’s continuity and precision. Notably, with 64 units, the discontinuities in the knee region were nearly eliminated, and the entire reconstruction process was completed in just 6.69 milliseconds. Moreover, adjustments in joint rotation angles effectively reduced reconstruction errors, bringing them down to 1.317 cm at a sampling frequency of 1/60. Additionally, iterative reconstruction decreased the error in the knee region by approximately 30% compared to non-iterative methods. The optimization process facilitated by the particle swarm algorithm revealed that most particles achieved high fitness levels after the initial iteration, and a considerable proportion shifted to the foreground region during the second iteration once fitness values dropped below 0.2. Ultimately, the study substantiates the accuracy and reliability of the proposed method for joint area reconstruction. These findings indicate that the FEA-based 3D reconstruction framework is highly effective in capturing complex joint movements and deformations, providing valuable insights for both biomechanical research and clinical applications.