Rule-based Image Segmentation Method for Continuous-tone, Text and Halftone in Electrophotographic Copier/Printer

摘要

In the area of digital copier/printer, when a document with mixed image types {continuous tone photographs, halftone pictures, text (gray as well as binary) and graphics} are used with one set of output printing algorithm at one time (error diffusion, gray level halftone screening, or thresholding etc.), there are some undesirable effects occurring. If an input halftone picture was printed with a lower screen frequency halftone output (such as a hard dot 150 lines/inch screen), moire pattern shows up. If a text image is printed with a halftone output, the text quality (especially gray text and smaller font size text as well) degrades due to the screening. If a more continuous tone like output algorithm (like gray level error diffusion or a partial dot higher screen gray level halftone) is used, then the granularity of the continuous tone input image suffers in an Electrophotographic system. An image segmentation method using feature extraction, fuzzy logic and rule based processing on scanned input images (at 600 dpi) to identify the image type in a pixel basis for printing is used. Two different output printing algorithms (a 300 l/inch partial dot screen for text and a mixed dot 161 l/inch screen output for continuous tone input) were used for different segmented image type in printing to achieve good result.