djb@wjh12.harvard.edu (David J. Birnbaum) (01/11/90)
In article <1990Jan11.040933.23748@uncecs.edu> ruslan@uncecs.edu (Robin C. LaPasha) writes: > Software: > PC: > SPOT (Flagstaff Engineering) > CAT > Recognita [plus](SzKI) > Datacopy > OCR Plus (Xerox) > (something by) (Calera Recognition Systems) > [in development: Textpert] > Kurzweil program...? > Mac: > Accutext (Xerox) > Textpert > OmniPage I have been using SPOT regularly for about a year and find it very satisfactory (although not error-free) for Cyrillic. Datacopy, like Kurzweil, is a division of Xerox. Datacopy makes OCR Plus. I used this product for a couple of months and then gave up on it; it is hardwired for some Latin alphabet "look-alike" checking that can not be disabled and that severely limits its flexibility for non-Latin alphabets. I found their technical support and customer relations policies unacceptable. This was over a year ago and both the soft- ware and personnel may have improved. Calera (formerly Palantir) makes TrueScan (a hardware expansion board with software), which is not trainable. Caere makes OmniPage (for MS-DOS; I don't know about MacIntosh), which is also not trainable. Textpert, by CTA, is in beta testing for MS-DOS; it has been available for the MacIntosh for a while. I will post a note when my article on non-Latin OCR appears in the Bits & Bytes Review. It will cover only trainable programs for MS- DOS, which is a fairly limited list. --David ============================================================ David J. Birnbaum djb@wjh12.harvard.edu [Internet] djb@harvunxw.bitnet [Bitnet] ============================================================