zernike

[ XITE Reference Manual | XITE home ]

Name

zernike - zernike moment image of a gray scale or binary image

Syntax

zernike [-n <n>] [-r] [-t <title>] <inimage> <outimage>

Description

The Zernike moment of order n and repetition m of an image f(x,y) is defined as follows:
          n+1                             *
 A(n,m) = ---- Sum Sum f(x,y)[V(n,m, x,y)]
           pi   x   y

 where x^2+y^2 <= 1

The image V(n,m, x,y) is the Zernike basis images of order n and repetition m. These basis images are complex and orthogonal. The Zernike moments are essentially the projections of the input image onto these basis images.

The original image can be reconstructed from the Zernike moments. The N-th order approximation is given by
  ^         N
  f(x,y) = Sum Sum A(n,m) V(n,m, x,y)
           n=0  m

The contribution or information content of the n-th order moments is

  I(x,y, n) = Sum A(n,m) V(n,m, x,y)
              m

Restrictions

inimage must be single-band with pixel type unsigned byte.

Options

-n n
Use moment order n. Default 0.

-r
Reconstruct image from moments. Default: Compute the absolute value of I(x,y, n).

-t title
Use title for outimage.

See also

zer_mom(3), zer_con(3), zer_rec(3), zer_pol(3)

References

[1] A. Khotanzad and Y.H. Hong
"Rotation invariant image recognition using features selected via a systematic method", Pattern Recognition, vol.23, no.10, pp.1089-1101, 1990.

[2] Thomas H. Reiss
"Recognizing Planar Objects Using Invariant Image Features", Lecture Notes in Computer Science, volume 676, pp. 17-20, Springer-Verlag, 1993.

Return value

0 : OK

Files

 xite/src/zernike/zernike.c

Author

Řivind Due Trier, Ifi, UiO.