Ngemva kweminyaka emithathu kusukela kwakhishwa okokugcina, isistimu yokuqaphela umbhalo ye-Ocrad 0.28 (Optical Character Recognition), eyathuthukiswa ngaphansi kwephrojekthi ye-GNU, isikhishiwe. I-Ocrad ingasetshenziswa kokubili ngendlela yomtapo wolwazi wokuhlanganisa imisebenzi ye-OCR kwezinye izinhlelo zokusebenza, futhi ngendlela yokusetshenziswa okuhlukile okuthi, ngokusekelwe esithombeni esidluliselwe kokokufakayo, ikhiqize umbhalo ngombhalo wekhodi we-UTF-8 noma 8-bit.
Ngokubonwa kokubona, i-Ocrad isebenzisa indlela yokukhipha isici. Kufaka phakathi isihlaziyi sesakhiwo sekhasi esikuvumela ukuthi uhlukanise kahle amakholomu namabhulokhi wombhalo kumadokhumenti aphrintiwe. Ukuqashelwa kusekelwa kuphela izinhlamvu ezisuka kumbhalo wekhodi we-"ascii", "iso-8859-9" kanye ne-"iso-8859-15" (akukho usekelo lwezinhlamvu zesiCyrillic).
Kuyaphawulwa ukuthi ukukhishwa okusha kufaka phakathi ingxenye enkulu yokulungiswa okuncane nokuthuthukiswa. Ushintsho olubaluleke kakhulu kwaba ukusekelwa kwefomethi yesithombe se-PNG, esetshenziswe kusetshenziswa umtapo wezincwadi we-libpng, okwenze kwaba lula kakhulu ukusebenza nohlelo, njengoba ngaphambilini bekungafakwa kuphela izithombe ezikumafomethi e-PNM.
Source: opennet.ru