An image noise reduction method in frequency domain is provided, comprising selecting a first image and applying a Fourier transform thereon to acquire a first frequency spectrum, wherein each pixel of the first frequency spectrum has a real part X and a imaginary part Y, calculating a first energy of each pixel and a first mean energy of all first energies of the pixels, calculating a first mean value and a first standard deviation of the real part of the pixels, calculating a second mean value and a second standard deviation of the imaginary part of the pixels, determining a first and a second predetermined values, when the first energy exceeds the first mean energy, the first and second predetermined value are replaced with X and Y, and when the first energy does not exceed the first mean energy, X and the Y are reserved.

 
Web www.patentalert.com

< Method of wavelet coding a mesh object

> Image processing apparatus for scaling images

> Dynamic range detection and adjustment

~ 00530