Opencv sift feature matching after how to calculate the similarity of two images? Online are introduced how to match the characteristics, did not say how similarity in the conclusion, Daniel, please advise, I have a small white

CodePudding user response:

Similarity is generally through the calculation of Euclidean distance,

CodePudding user response:

Their top, temporarily suspended

CodePudding user response:

No prawns out his top, don't sink

CodePudding user response:

http://blog.csdn.net/wangyaninglm/article/details/43853435 this?

CodePudding user response:

The size of the commonly used Euclidean distance comparison, specific can refer to my blog:
SIFT algorithm of feature point selection and DMatch, Keypoint description

CodePudding user response:

reference 5 floor codedoctor reply:

general using Euclidean distance comparison, the size of the concrete can refer to my blog:
SIFT algorithm of feature point selection and DMatch, Keypoint description

Article is no difference between the two images of similarity? I tested two completely different figure, can also be a lot of feature point matching

CodePudding user response:

reference 4 floor iloveyou418 response:

http://blog.csdn.net/wangyaninglm/article/details/43853435 this?

I also tried, this article PSNR didn't compile, a pile of gpu related definition of error, don't know what less? Several other methods effect is not very ideal

CodePudding user response:

Or "TensorFlow"?

CodePudding user response:

refer to 7th floor u012733720 response:

Quote: refer to 4th floor iloveyou418 response:

http://blog.csdn.net/wangyaninglm/article/details/43853435 this?

I also tried, this article PSNR didn't compile, a pile of gpu related definition of error, don't know what less? Several other methods effect is not very ideal

Don't is import CV library?

CodePudding user response:

All things can train,
- ward basis, to realize

CodePudding user response:

refer to 6th floor u012733720 response:

Quote: refer to the fifth floor codedoctor reply:

The size of the commonly used Euclidean distance comparison, specific can refer to my blog:
SIFT algorithm of feature point selection and DMatch, Keypoint description

Article is no difference between the two images of similarity? I tested two completely different figure, also can match a lot of feature points on

Judging by Euclidean distance, Euclidean distance is too large is certainly not a figure

CodePudding user response:

The building Lord found the solution

CodePudding user response:

The building Lord, please find the answer? I also want to have two images using SIFT the similarity calculation and numerical value is given