Platform: win10 x64 + VS 2015 Pro + opencv-2.4.11 + gtk_-bundle_2.24.10_win32
Main reference: 1 Code:. RobHess source of SIFT
2. Book: Jin Gui Yongming "invariant local image features with the description"
SIFT four steps:
Step one: Create scale-space, namely the establishment of Difference of Gaussian (DoG) pyramid dog_pyr
Step two: detecting extreme points, precise positioning and size and filter to create a default scale space in the memory storage
Step 3: assign a feature point direction, to complete this step, there are three of each feature point information: location, size, direction
Step Four: calculating feature descriptor
Step Four: calculating feature descriptor
Tri-linear interpolation:
. 1) the SIFT Algorithm: feature descriptor : https: //www.cnblogs.com/JiePro/p/sift_4.html
2) Description Vector: https: //blog.csdn.net/zddblog/article/details/7521424
3) 4.3 trilinear interpolation: https: //www.sun11.me/blog/2016/sift-implementation-in-matlab/
4) Why were tri-linear interpolation? https://blog.csdn.net/fzthao/article/details/62424271
5) I continued with the interpolation meet by chance (3): tri-linear interpolation (Trilinear Interpolation) principle and use: https: //blog.csdn.net/Xyc19930930/article/details/77806504
6) trilinear interpolation (three-dimensional linear interpolation): https: //blog.csdn.net/webzhuce/article/details/86585489
7) linear & bilinear interpolation & trilinear interpolation: https: //blog.csdn.net/sinat_39372048/article/details/81361038