Study On Halftone Image Descreening

With the fast development of electronic publication, some existed printed images often need to be scanned. The printed image is produced by halftone technique, with different kinds of screen information, so it need to descreen through scanning. Halftone image descreening, that is the very important step to achieve printed image digitization, has the very important research meaning and apply value.Firstly, in the subject, error diffused halftone image is implemented and analyzed its character .Based on it, we study and implement seven kinds of smooth lowpass filters for descreening. After experiment and compare of objective evaluated standard, find the lowpass filter based on HVS function, Gaussian lowpass filter and FIR lowpass filter get good results. After that a lowpass filter based on edge enhancement is used and get a better result. For the scanned halftone image, we use two-dimensional FFT transform frequency analysis to detect the presence and frequency of screen, and then choose Gaussian lowpass filters and suitable coefficient. Then use wavelet transform for halftone images, through the research analysis, and find the screen lines and screen angle are not all same in different component in CMYK color space. So choose different filters along specific orientations in C, M, Y, K components and get a better result. In the performance analysis of the algorithm, image quality is measured objectively using Image definition, PSNR (peak signal-to-noise ratio), SNR (signal-to-noise ratio) and so on. And get the same result with subjective evaluationAfter analyzing the experiment results, show the method get good effect and practicability, make the certain foundation for the further research.