OCR recogniton (Optical Character Recognition) extracts and analyzes the shape of an image associating it to a character value. Such value is determined on the basis of a complex mathematical analysis, together with template criteria of matching, extracting and analyzing images features. This analysis produces a series of possible results, each with its own level of confidence. Each OCR recogniton engine uses its own methodology. That's why Recogniform OCR Voting System acts, integrated in Recogniform OCR Server, processing the same images with different OCR engines, analyzing for every character the various possible alternatives and choosing the best solution for each character, reducing the possibility of errors. Recogniform OCR Server is able to process in completely AUTOMATIC way a UNLIMITED number of documents, of every form/kind/size. Through the monitoring of a folder in INPUT, Recogniform OCR Server processes all files contained in the input folder and produces the output desired, in the chosen format. Recogniform OCR Server executes at first all the preliminary operations on the image (deskew, despeckle, cleaning), then proceeds with the layout analysis (the segmentation in lines/columns/blocks of text/images, etc.) and at last performs the OCR recognition on the page, releasing the output in PDF, PDF-A, .TXT or .RTF format, in the user-predefined output directory.
Practically, once the processing starts, Recogniform OCR Server works in total autonomy, without no necessity of human activity. Currently Recogniform OCR Server is used by some important companies performing massive document OCR processing.