获取Matrix来训练sift特征的问题
我已经使用 OpenCV 开源提取了特征。
我已经使用这两个函数完成了这些步骤
SiftFeatureDetector
SiftDescriptorExtractor
，我从描述符中得到了 128*128 的矩阵，我也认为我将使用它 这个矩阵来训练特征...
我感到困惑的是，
当我想训练特征时， 我应该使用一个特征数量矩阵，每一行都包含有关该特征的信息。它可能是一个矩阵，
number of features * 6
例如，我在图像中得到了 344 个特征......我得到了一个 128*128 的矩阵对于描述符，我需要这个矩阵来训练我的特征
，但正如我提到的，我只是得到 128*128 矩阵..那么问题是什么？
还有，以后我应该训练什么？
I have extracted the features by using OpenCV opensource..
I have done these steps by using these 2 functions
SiftFeatureDetector
SiftDescriptorExtractor
which I got a matrix of 128*128 from the descriptors, which I think as well that I will use
this matrix to train the features...
What I'm confused about is the following,
When I want to train the features,
I should use a matrix of number of features and every single row contains the information about that feature.. which it might be a matrix of
number of features * 6
For example, I got 344 features in an image... and I got a matrix of 128*128 for the descriptor, which I need this matrix in order to train my features
but as I mentioned, I'm just getting 128*128 matrix.. so what's the problem?
And, what should I get to train later on?
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（1）
您是否看过descriptor_extractor_matcher.cpp，或 matcher_simple.cpp 来自 OpenCV 的示例？另外，您可以发布您用来检测功能的代码吗？
Have you looked at the descriptor_extractor_matcher.cpp, or the matcher_simple.cpp samples from OpenCV? Also, could you post the code you are using to detect the features?