David의 개발 이야기!

[CV Lecture 10] - Image Alignment 본문

컴퓨터비전

[CV Lecture 10] - Image Alignment

david.kim2028 2023. 12. 17. 20:33
반응형

A.  Least Squares

Suppose we need an affine transformation to align one image to the other one.

 

 

 

 

 

위 식을 정리를 하면 아래처럼 표현할 수 있다.

 

 

 

정리된 식을 미분하면, 세타 에 대한 식이 나오는데, 여기 양변에 A 를 곱해주면, 

 

A * theta = b 

 

식이 나오게 된다. 

 

따라서, 해당 문제를 푸는것이 문제가 된다.

 

 

 

즉, image alignment 를 위해서, A, B에서 interest point 를 찾고, 이를 이으면서, homography 를 추정한다.

homography 행렬은, 한 이미지의 관심점들을 다른 이미지의 관심점으로 변환하는데 사용할 수 있으며, 두 이미지간의 변환을 구현할 수 있다.

 

 

B.  RANSAC

: Random SAmple Consensus

 

1. 알고리즘

- 1. Sample randomly the number of points required to fit the model

- 2. Solve for model parameters using samples

- 3. Score by the fraction of inliers within a preset threshold of the model

- 4. Repeat 1-3 until the best model is found with high confidence

 

 

2. Ransac 디테일적인 부분들

* bestmodel = inliers 임 (오타)

 

 

 

오늘은 image align 하는 기초와 그 대표적인 알고리즘인 ransac 에 대해 공부했다!

반응형
Comments