平滑骨架数据（自定义），预处理 [英] Smoothing Skeleton Data (custom), Pre-Processing

问题描述

您好

我只是想知道是否有人知道可以在线研究的任何方法和/或在序列化之前平滑骨架数据的任何教程联合数据？在自定义算法或预测算法的意义上。

I am just wondering if anyone knows any methods that can be researched on-line and/or any tutorial on smoothing skeletal data before serialization of the joint data? In the sense of custom algorithms or predictive algorithms.

我 已经使用
TransformSmoothParameters 已经帮助了学位但仍然不够好的课程。

I have already used the TransformSmoothParameters class which has helped to a degree but still not good enough.

对于我正在做的项目而言，处理速度并不重要，如果它滞后并不重要很多，它不适合在游戏中使用。

With the project i am doing there is no importance of speed of processing and it does not matter if it lags a lot, it is not for use in a game.

我已经能够实现一些后处理以数学方式纠正数据但出于准确性的原因我想尽量减少数量输入错误。

I have been able to implement some post-processing to correct the data mathematically but for accuracy reasons i would like to minimise the amount of errors at input.

如果这是一个简单的问题，请道歉但是我通常是一个UI类型，这对我来说相当科学！

Apologies if this is a bit of a simple question but I'm normally a UI type, this is fairly scientific for me!

问候，

David 

David 

推荐答案

为什么不关闭平滑参数，然后应用算法？ 使用基于深度数据的每像素分析的决策树来确定关节位置，因此在没有平滑的情况下生成的原始关节
之前可能没有太多可以做的。

Why not just switch off the smoothing parameters, and then apply your algorithms?  The joint locations are determined using a decision tree based on a per-pixel analysis of the depth data, so there probably isn't much you can do prior to the raw joints produced without smoothing.

这篇关于平滑骨架数据（自定义），预处理的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！