Program Codes

This page contains source programs coded by our laboratory. They are formatted by ``tar'' and ``gzip''. If you download them, unpack them by ``gunzip'' and ``tar'' and follow the instructions described in the README files. All the programs are in the public domain; they do not contain licensed codes or parts of commercial software.

Geometric Computation

Fundamental matrix computation (C++) (2003/6/16 correct.c fixed)

This program optimally computes the fundamental matrix from point correspondences between two images. When you unpack the package, a directory named ``fundamental'' is automatically created. For the details, see README, which will be created in that directory.

References:

K. Kanatani, Optimal fundmamental matrix computation: Algorithm and reliability analysis(.ps.gz),
Proc. 6th Symp. Sensing via Image Information (SSII 2000), June 2000, Yokohama, Japan, pp. 291--298.

Flow fundamental matrix computation (C++) ( zip for windows)

This program optimally computes the flow fundamental matrix from point correspondences between two images regarded as defining optical flow. When you unpack the package, a directory named ``flowmat'' is automatically created. For the details, see README, which will be created in that directory.

References:

K. Kanatani, Y. Shimizu, N. Ohta, M. J. Brooks, W. Chojnacki and A. van der Hengel, Fundamental matrix from optical flow: Optimal computation and reliability evaluation,(.ps.gz),
Journal of Electronic Imaging, Vol. 19, No. 2 (2000), 194--202.

Homography computation (C++)

This program optimally computes the homography from point correspondences between two images. When you unpack the package, a directory named ``Homography'' is automatically created. For the details, see README, which will be created in that directory. To compile this program, you need the library file ``matrix''. Download it and set it up before compiling the homography program.

References:

K. Kanatani and N. Ohta, Accuracy bounds and optimal computation of homography for image mosaicing applications(.ps.gz),
Proceedings of the 7th International Conference on Computer Vision, 20--27 September 1999, Kerkyra, Greece, pp. 73--78.

3-D rotation estimation (C++) (2010/5/25 Corrected)

This program optimally estimates a 3-D rotation matrix from two sets of corresponding 3-D points. When you unpack the package, a directory named "Rotation" is automatically created. For the details, see README, which will be created in that directory. To compile this program, you need the library file ``matrix''. Download it and set it up before compiling the rotation program.

References:

N. Ohta and K. Kanatani, Optimal estimation of three-dimensional rotation and reliability evaluation(.ps.gz),
Proc. 5th Euro. Conf. Computer Vision (ECCV'98), June, 1998, Freiburg, Germany, Vol. 1, pp. 175-187.

Conic fitting (C++)

This program optimally fits a conic to given points in two dimensions. When you unpack the package, a directory named ``conicest'' is automatically created. For the details, see README, which will be created in that directory.

References:

Y. Kanazawa and K. Kanatani, Optimal conic fitting and reliability evaluation(.ps.gz),
IEICE Transactions on Information and Systems, Vol. E79-D, No. 9 (1996), pp. 1323--1328.

Image Processing

Background Subtraction (C)

This program executes background subtraction robust to illumination chage. When you unpack the package, a directory named "ImDiff" is automatically created. For the details, see README, which will be created in that directory.

References:

N. Ohta, A Statistical Approach to Background Subtraction for Surveillance Systems (.ps.gz)
Proc. of ICCV 2001, pp. 481-486, Vancouver, Canada.

Libraries

Matrix (C++): This library file contains various tools for vector/matrix/tensor computation. When you unpack the package, a directory named "Matrix" is automatically created. For the details, see README, which will be created in that directory.