Tracking using Pixel-Wise Posteriors

最新推荐文章于 2024-11-13 01:30:20 发布

weixin_30644369

最新推荐文章于 2024-11-13 01:30:20 发布

阅读量214

点赞数

CC 4.0 BY-SA版权

原文链接：http://www.cnblogs.com/CVArt/archive/2011/10/19/2217539.html

本文介绍了一种使用像素级后验处理的鲁棒实时视觉跟踪方法，该方法通过帧间的刚性注册、残余形状变形的分割和在线外观学习实现了对象和背景外观模型的连续优化。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

　　之前了解过一些效果较好的跟踪算法，如跟踪人手的HandVu所用的黄金金字塔跟踪、号称可以和三维跟踪媲美的TLD跟踪算法，但现在这个跟踪算法不由为之一振，

只能说牛人太多了啊= =！

The tracking problem is handled using a bag-of-pixels representation and comprises a rigid registration between frames, a segmentation and online appearance learning. The registration compensates for rigid motion, segmentation models any residual shape deformation and the online appearance learning provides continual refinement of both the object and background appearance models. The key to the success of our method is the use of pixel-wise posteriors, as opposed to likelihoods.

项目主页

Robust Real-Time Visual Tracking using Pixel-Wise Posteriors – Paper(Published ECC08)
Robust Real-Time Visual Tracking using Pixel-Wise Posteriors – Presentation (Published ECC08)

转载于:https://www.cnblogs.com/CVArt/archive/2011/10/19/2217539.html