Research On Inverse Halftoning Method Based On Look-up Table

Image half-toning technology reduces reappearance cost of imageoutput. Meanwhile, according to different half-toning algorithms, thenoise of different degrees is added to the image, which decreases imagequality. If reinforcement, compression and digitization are conducted forthese half-toned images, it is required to reconstruct and restore them, i.e.adverse half-toning processing. Currently, half-toning and adversehalf-toning techniques of grayscale images are relatively mature.However, as digital devices prevail, color image processing becomes themainstream processing object. In numerous adverse half-toning methods,look-up table is a machine learning method with good processing effectsand small storage space which can be easily achieved and does notdepend on any linear filter. This paper mainly studies image adversehalf-toning method based on look-up table as follows:1) Image adverse half-toning method based on classical look-up tableis optimized. In the optimization algorithm, optimized training templatewas selected.“Null value” valuation method is improved. Aiming at theshortcomings of “null value” valuation method, eight-neighborhooddynamic mean value estimation method is put forward. Thus, the qualityof adverse half-toned image is improved.2) Image adverse half-toning method based on sorted color storing isproposed. This method changes the form of single-way table creating intraditional methods, comes up with the idea of both-way table creatingand stores the continuous values with many noises in the alternate table asimportant basis of “null value” valuation. Visual effect of color adversehalf-toned image gained through this method is good. Besides, the meanvalue of evaluation indexes is high.3) Quality evaluation method for adverse half-toned images–RGBcomponent visual evaluation component is put forward. This evaluationmethod is composed of color rendition degree and color renditiontendency. Color rendition degree reflects the color difference degreebetween adverse half-toned images and continuous images. Color rendition tendency reflects tendency direction and degree of colorrendition of adverse half-toned images. The values of the two indexes arevisual and pellucid and comply with human eye vision.Image half-toning method based on look-up table takes full advantage ofthe features of continuous images and priori knowledge and reducesinformation loss and distortion degree of adverse half-toned images. Inaddition, the processing speed is also superior to other adversehalf-toning algorithms. Thus, this method owns higher follow-up researchvalue.