OpenCV手势识别的现状？
OpenCV 中手势识别的现状如何？
我见过能够检测手势的很好的例子（例如 https://www.andol.me/1661 /），但识别手势并执行某些操作（例如操纵屏幕上的对象）似乎要困难得多。
有谁知道有什么例子吗？
谢谢！
What is the current state of hand gesture recognition in OpenCV?
I have seen great examples of being able to detect hand gestures (e.g. https://www.andol.me/1661/) but recognising gestures and performing some action (e.g. manipulating on-screen objects) seems much harder.
Does anyone know of any examples?
Thanks!
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（2）
我认为 OpenCV 对手势识别的唯一支持正是 Luca Del Tongo 在您链接的视频中演示的，即 cvConvexityDefects()。
您可能想要使用颜色空间过滤（如视频中建议的那样）提取手部遮罩，并使用 ML 分类器来检测实际手势。有几篇论文已经做到了这一点，尽管目前 OpenCV 中还没有这样实现。
I think the only support in OpenCV for hand gesture recognition is exactly what Luca Del Tongo demonstrated in the video you linked to, namely cvConvexityDefects().
You might want to extract the hand mask using color-space filtering (as suggested in the video), and using ML classifiers for detecting the actual gesture. There have been several papers that have done this, though this is not currently implemented as-such in OpenCV.
cvConvexityDefects() 是手部分割之后的关键步骤，如 andol.info/hci/1661.htm 中所做的那样
cvConvexityDefects() is the key step following the hand segmentation as done in andol.info/hci/1661.htm