基于双学习映射增量降维模型的人体运动形态估计

如何从二维的人体运动形态图像序列估计出三维的人体运动形态是一个热门的研究课题。这个项技术的实现需要建立复杂模型,设计复杂的训练算法,并且需要大量的人体运动形态样本进行训练。然而,利用目前的一些复杂模型和算法,估计三维人体运动形态会耗费很多时间,并且估计效果也不好。因此,要想得到更准确的估计结果还是比较困难的。对于这样的一个研究问题,提出基于双学习映射增量降维模型来实现三维人体运动形态估计,该模型可以从单视角的二维图像较好的估计出三维人体运动形态,缩短估计算法运行时间,并得到更准确的估计结果。

Dual Learning Incremental Dimension Reduction Model for Human Motion Estimation

Estimating 3 dimensional(3D) human motion from image sequence is a hot research topic. This technique requires a lot of complex algorithms and models, and needs numerous samples to train. Furthermore, the 3D human motion costs a lot time to be estimated, when recent some complex algorithms and models are utilized. Thus, it is difficult to achieve more accurate results of the estimation. For such a problem, a new method called dual learning incremental dimension reduction model (DLIDRM) is proposed to achieve the 3D human motion estimation, it can estimate the 3D human motion more efficiently from single view, save the running time and get more accurate results.