Readme.md

Image Alignment Methods included in this package:

1. Single-Image Alignment
   • Ellipse-based: Align an image so that its main axes are the y and x axis of the cartesian coordinate system. Rotates an image so that it is straight up-and-down.
     Gwo C and Wei C (2016). “Shoeprint Retrieval: Core Point Alignment, for Pattern Comparison.” Science & Justice, 56(5), pp., 341-350. ISSN 1355-0306, doi: 10.1016/j.scijus.2016.06.004 (URL:, http://doi.org/10.1016/j.scijus.2016.06.004), 00000, <URL:, http://www.sciencedirect.com/science/article/pii/S1355030616300521>.
2. Dual-Image Alignment
   • Keypoint Alignment: Keypoint detection using the following steps:
     1. Harris corner detector
     2. Calculate key image orientations
     3. Feature detection
     4. Point matching using KNN to find similar points
     5. RANSAC to get a consistent homology
     6. Image transformation