In Japan, many health care providers evaluate the recovery status of patients by observing a change in the patient's manner of walking. In the initial stage of rehabilitation, the patient's manner of walking is unstable. As the rehabilitation progresses, the manner of walking of the patients becomes stable. Therefore, changes in the foot print images of the patients may be detected. The techniques of noise-reduction, binarization of a foot print image, and erosion and dilation to smooth the edge of the binary image to detect the edge of the foot print image are described. In addition to these, discrete Fourier transformation to measure the similarity of the foot print patterns and inverse discrete Fourier transformation to remove high frequency components in the foot print pattern are described. Using these basic techniques, a method to estimate gait condition is examined.