Fourier coefficients for pattern analysis

32 Views Asked by At

There are many areas like, gait analysis, where we recognize persons by analyzing their silhouettes taken while they are at different stages of their walking where analysis also carried on in frequency domain . I want to know what exactly is the central idea for recognition in such algorithms . I guess a 2-D Fourier transform of the matrix which is obtained by subtracting the input image matrix from the image matrix of the subject ,with which the comparison is to be carried and then the power spectral density is computed by summing up the second power of all the coefficients . The lesser the P.S.D , the closer they are . Is my idea correct . I would like to know what exactly serves as the basis of comparison and recognition in frequency domain .