本文共 1135 字，大约阅读时间需要 3 分钟。

image alignment tools

SIFT keypoints and RANSAC algorithm(OpenCV library)

https://github.com/khufkens/align_images

Method：

1. FFT phase correlation 相位相关算法实现

（仅平移）

1. Enhanced Correlation Coefficient (ECC) maximization 增强相关系数最大化图像匹配

the ECC methodology can compensate for both shifts, shifts + rotations (euclidean), shifts + rotation + shear (affine), or  (3D) transformations of one image to the next.

1. Feature based registration 基于特征 （SIFT）

 uses features within the image rather than a correlation based method on the whole image to search for these values

Feature based registration （SIFT+SURF）

P.S.

feature detector: SIFT，SURF，ORB

——>https://learnopencv.com/feature-based-image-alignment-using-opencv-c-python/

1. AKAZE local feature detection
2. RANSAC

(with low efficiency)

https://github.com/vjayd/Image-Alignment-using-CNN

（CNN without data）

P.S.

Othor projects based python on github are  mostly related with star points, face alignment, OCR, and Imagefusion.

This is a demo of detecting and aligning faces in an image. Alignment is a process of rotating a face to a vertically straight orientation, should the original face image is tilted.

转载地址：http://lxtgn.baihongyu.com/