SIFT

SIFT (Scale-Invariant Feature Transform) is a computer vision algorithm that detects, describes, and matches distinctive image features, ensuring robustness against scale, rotation, and illumination changes.

SIFT, developed by David Lowe, is the gold standard for robust feature extraction in computer vision. The process involves four key stages: first, detecting scale-space extrema using the Difference of Gaussians; second, precisely localizing keypoints by fitting a 3D quadratic function; third, assigning a consistent orientation to each keypoint; and finally, generating a unique 128-element descriptor vector. This descriptor allows for highly reliable feature matching, enabling critical applications like 3D reconstruction, panoramic image stitching, and precise object recognition across varied scenes.

https://docs.opencv.org/4.x/d5/d51/tutorial_py_sift_intro.html