Fourier Image Filtering

On the left is the original (or filtered) image on the right the 2D FFT image. It shows all the frequencies in the image arranged with the lowest frequency in the center increasing in frequency away from this point towards the edge of the image. Therefore if you mask out the central bit of the FFT image you will lose this low frequency information in the filtered image, effectively doing a high pass filtering, if you only keep this central region then you will pass the low frequencies (low pass filtered image). If you consider the FFT image now as a compass, the direction is related to the 2D FFT orientation so by masking out pie shaped regions of the FFT image only frequencies of a certain orientation will be passed in the filtering. To perform the filtering mask out some of the image and click the iFFT button (if you dont mask the image nothing will be filtered and the image will look the same).

To filter an image first upload the image, the tool performs an automatic colour 2D FFT which is shown on the image on the right. The controls under the images allow different areas of the FFT image to be masked, by selecting the invert checkbox the mask undoes the masking of other tools. The angle of the arc tools can be adjusted to select only certain orientations. By clicking on the delete icon all drawing objects are removed from the FFT image. By clicking on the black tool the entire FFT image is filled. An iFFT is preformed by clicking the iFFT button and if any of the FFT image has been masked the image will be filtered accordingly. At the moment only hard edged tools have been implemented so therefore you can observe some ringing artefacts, at a later date I will introduce a Guassian-like tool to minimise this.