Appeal No. 1998-0671 Application 08/285,328 Algorithms that convert gray images to binary or other number of gray levels attempting to preserve gray density exist, and include among them error diffusion . . . . See also specification, p. 2, last paragraph, which essentially repeats the first paragraph above. Thus, Eschbach '653 discloses that a problem with screening is that the amount of gray within an original image is not maintained exactly over an area and that throwing away the error caused by thresholding causes a loss of image information and that it was known to use error diffusion to preserve gray density. There is no dispute that error diffusion performs the steps of "(d) generating an error value as a result of the reduction process . . ." and "(e) diffusing the error value to multi-level grey scale pixel values of adjacent pixels." The Survey article also discloses that error diffusion would be used to provide gray scale rendition where there is coarse gray scale quantization (p. 298): "The fundamental strategy is simply that of direct spatial distribution of the errors created by coarse gray scale quantization, and it can be applied to two or more gray level marking processes." The Examiner did not rely on this teaching of the Survey article. In our opinion, it would have been obvious to one of ordinary skill in the art to use error diffusion to distribute the error - 17 -Page: Previous 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 NextLast modified: November 3, 2007