Digital image processing is rapidly developing, and its use is found in various aspects of human activity, from science to engineering area that includes graphics technology. Modern digital technologies have enabled manipulation with parts of the image or the entire image in order of their improvement or adjustments to current needs. In this thesis the results of images filtered with filters for sharpening and smoothing in the spatial and frequency domain are experimentally compared to the original images. Results demonstrate influence of filters on the quality of the images. In addition, the speed of filtering was also measured, and the results were presented in tables and diagrams. The experimental procedure was performed on the database of greater number of grayscale images. Images were filtered with Unsharp and Gaussin filter in both spatial domain and frequency domain. After that the filtered images were compared to the originals. As a measure for the quality of the images the peak signal to noise ratio (PSNR) is used and expressed in decibels (dB). Filtering speed is measured in seconds. Results of the experiment have shown that the filtered images in the spatial domain have more quality than images filtered with filters in the frequency domain. The research has also shown that the speed of filtration in spatial domain is considerably greater when compared to the filtering speed in frequency domain.