基于遗传优化的自适应凸松弛人体姿势估计*

针对凸松弛方法在解决三维人体姿势估计的问题时存在迭代次数较多、准确度不高的不足，提出一种基于遗传优化的自适应凸松弛人体姿势估计算法。该算法首先对关键参数的更新方式进行自适应处理，然后利用遗传优化算法对该关键参数的初始值进行寻优，最后利用寻优结果对凸松弛方法中闭式解的公式进行调整。实验结果表明，提出的算法迭代次数更少，准确度更高，更有利于实际应用。

Human Pose Estimation: An Adaptive Convex Relaxation Approach Based on Genetic Optimization

In order to overcome the drawbacks of large number of iterations and low accuracy of convex relaxation approach during solving 3D human pose estimation, this paper proposed an adaptive convex relaxation approach based on genetic optimization. Firstly, it adaptively updated the key parameter. Then it used the Genetic Algorithm to optimize initial value of the key parameter. Finally it improved the expression of closed-form solution in the convex relaxation approach by using the result of optimization. The experimental results show that the improved algorithm has less iterative numbers but higher accuracy, which could be much helpful to practical applications.