推断联合算法 [英] Inferred Joint Algorithm
问题描述
我对kinect估计联合位置时有点好奇,因为它是INFERRED。它使用什么算法?
有几个原因可以将联合的跟踪状态报告为推断。通常,这意味着算法估计位置的置信度不符合最小阈值。这可能是由于关节被隐藏在视图
(例如,由另一个身体部位)或在摄像机的视野之外(太高或太低,或偏离侧面)。这也可能是由于深度数据过于模糊,导致算法无法在关节的两个或多个可能的
位置之间进行明确选择。
如果你微软研究院的这篇论文解释说:'对这个算法是如何工作感兴趣的:
http:// research.microsoft.com/apps/pubs/?id=145347
John
K4W Dev
Hi,
i am a little bit curious about how kinect estimate joint position when it is INFERRED. What algorithm it uses?
There are several reasons that the tracking state for a joint may be reported as inferred. Generally, it means that the algorithm's confidence in estimating the position did not meet a minimum threshold. This may be due to the joint being hidden from view (e.g., by another body part) or outside the camera's field of view (too high or too low, or off to the side). It may also result from too much ambiguity in the depth data, causing the algorithm to not be able to definitively choose between two or more possible positions for the joint.
If you're interested in how the algorithm works, this paper by Microsoft Research explains it: http://research.microsoft.com/apps/pubs/?id=145347
John
K4W Dev
这篇关于推断联合算法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!