VLFeat is a C library, that you can use from C++ (without changing the library). The images have to be stored as one float per pixel. You can read the C API documentation here: http://vlfeat.org/api/sift_8h.html
You can also checkout Rob Hess's SIFT implementation. The implementation is done step by step as mentioned in David Lowe's SIFT paper. Quite useful for those who want to understand SIFT.
发布评论
评论（2）
VLFeat 是一个 C 库，您可以在 C++ 中使用它（无需更改库）。图像必须存储为每个像素一个
float。您可以在此处阅读 C API 文档：http://vlfeat.org/api/sift_8h.htmlVLFeat is a C library, that you can use from C++ (without changing the library). The images have to be stored as one
floatper pixel. You can read the C API documentation here: http://vlfeat.org/api/sift_8h.html您还可以查看Rob Hess 的 SIFT 实现。正如 David Lowe 的 SIFT 论文中提到的，该实现是逐步完成的。对于那些想了解 SIFT 的人来说非常有用。
VLFeat 和 VLFeat 都Hess 的软件是屡获殊荣的软件。
Utkarsh Sinha 在 SIFT 上创建了一个很好的教程。
You can also checkout Rob Hess's SIFT implementation. The implementation is done step by step as mentioned in David Lowe's SIFT paper. Quite useful for those who want to understand SIFT.
Both VLFeat & Hess's are award winning software.
Utkarsh Sinha has created a nice tutorial on SIFT.