Message-Id: <mailto:199408270307.WAA11039@library.wustl.edu> Date: Fri, 26 Aug 1994 15:08:22 -0400 From: Arthur Matheny <mailto:matheny@RITCHIE.ACOMP.USF.EDU> Subject: Re: Image database/ RE: Huffman encode (FAX-style) encryption and To: Multiple recipients of list IMAGELIB <mailto:IMAGELIB@ARIZVM1.BITNET>
On Fri, 26 Aug 1994, Raymond Lauzzana wrote:> When you speak of Type III and Type IV compression, what categorization are
> you speaking of. Has some organization established standards, like ISO?
> What we have is numerous catalogs. most pages are simply text, many are non-Engl
> ish
> Sometimes a photograph occures, but but that is infrequent in comparizson to
> the overall document.
>
> What I am considering is Huffman encoding the entire document, and havingg
> an association pointer with positioning ing info connecting a PICT or TIFF
> file to pages in which pictures occur. This strategy could produce much
> more efficient file sizes. It just doesn't seem appropriate to encode the
> entire page as a picture when most of it is text.
As I understand it, CCITT Group 3 or Group 4 compression can _only_ be applied to single-bit-plane images. However, it seems to me that gray scale pictures can be converted to dithered black & white, if that is acceptable quality. The advantage would be that you could store the pages the same way without regard to the presence of pictures. The compression might do poorly on the pictures, but if your documents are mostly text, it ought to do very well on average.
------------------------------------------------------------------ Arthur Matheny Numerically Intensive Computing Consultant Academic Computing LIB 612, University of South Florida 813-974-1795 Tampa, FL 33620 ------------------------------------------------------------------