readme

sift的matlab代码

                         AN IMPLEMENTATION OF
               LOWE'S SCALE INVARIANT FEATURE TRANSFORM

                            Andrea Vedaldi
                         [e-mail]
                   http://www.cs.ucla.edu/~vedaldi/

This is a MATLAB/C implementation of the SIFT detector and descriptor
[1]. See the documentation in 'doc/' to get started.

COMPLING
  From MATLAB prompt
  > sift_compile
  From shell prompt (Linux, Mac OS X)
  > make

TESTING
  > sift_demo
  > sift_demo2
  > sift_demo3

ABOUT THE SOURCE CODE
  We use the follwing convention to name files:
  *.m *.c:       M and MEX files
  sift*mx.c:     `Private' MEX files 
  sift_*.m:      Service scripts (compiling, demos)

CHANGES

  0.9.12  Removed LAPACK call. Added LOWE_BUG symbol definition to
          reproduce a small issue with Lowe's implementation in the
          computation of the keypoint orientations and match its
          output more closely.
  0.9.11  Corrected a bug which prevented Magnif to function properly.
          Removed a spurious KEYBOARD command.         
  0.9.10  Added sift_demo6.m.
  0.9.9   Added a flag to compile on Intel Macs.
          Improved the documentation.
  0.9.8   Minor bug fixes
  0.9.7   Fixed a minor glitch in the convolution code. Increased the
          default number of octaves that get computed (this is more
          similar to Lowe's version).
  0.9.6   Modified the convolution code to handle image boundaries more
          similarly to Lowe's version.
  0.9.5   Fixed a minor issues with the new sift_demo3.m.
  0.9.4   Fixed another bug in the detection of keypoints close to
          the boundary. Reduced the size of the distribution. New test code.
  0.9.3   Improved plotmatches.m and tightsubplot.m
  0.9.2   Improved detection of keypoints too close to the boundary
  0.9.1   Fixed a bug in the creation of the .zip archive
  0.9     Fixed a bug for which some of the keypoint detected where
          dropped.

CREDITS
  The SIFT algorithm [1] has been patented by David Lowe. Some of the
  images in 'data/' are from [2,3].

[1] D. G. Lowe, "Distinctive image features from scale-invariant
    keypoints," IJCV, vol. 2, no. 60, pp. 91 110, 2004.

[2] K. Mikolajczyk, T. Tuytelaars, C. Schmid, A. Zisserman, J. Matas,
    F. Schaffalitzky, T. Kadir, and L. Van Gool, "A comparison of affine
    region detectors," IJCV, vol. 1, no. 60, pp. 63 86, 2004.

[3] C. Hormann, "Landscape of the week 2," 2006.