ABBYY FINEREADER EXPRESS 8.3 MANUAL
A special feature of our process model is that manual interventions lead to the production of high quality GT being used as additional training data, thus enabling a spiral towards continuously higher automation. Our focus is throughout on an easy-to-use and efficient method, employing automatic methods where feasible and resorting to manual intervention where necessary. OCR4all covers all steps of an OCR workflow from preprocessing, document analysis (segmentation of text and non-text regions on a page), model training, to character recognition of the text regions.
ABBYY FINEREADER EXPRESS 8.3 SOFTWARE
The present paper describes our efforts to collect these recent advances into an easy-to-use and platform independent software environment called OCR4all that enables an interested party to obtain a textual digital representation of the contents of these printings. For a successful supervised training process, the Ground Truth (GT) in the form of line images and their corresponding transcriptions has to be manually prepared as training examples. While modern fonts can be recognized with excellent accuracy by so-called omnifont or polyfont models, very early printings like incunabula (books printed before 1501), but also handwritten texts usually require book-specific training in order to reach Character Error Rates (CERs) well below 10% or even 5%, as shown by Springmann et al. Furthermore, the non-standardized typography represents a big challenge for OCR approaches. Among the problems that need to be addressed for early printings is the often intricate layout containing images, ornaments, marginal notes, and swash capitals. While Optical Character Recognition (OCR) is regularly considered to be a solved problem, gathering the textual content of historical printings using OCR can still be a very challenging and cumbersome task, due to various reasons. The architecture of OCR4all allows the easy integration (or substitution) of newly developed tools for its main components by standardized interfaces like PageXML, thus aiming at continual higher automation for historical printings.
Furthermore, on very complex early printed books, even users with minimal or no experience were able to capture the text with manageable effort and great quality, achieving excellent Character Error Rates (CERs) below 0.5%. During experiments, the fully automated application on 19th Century novels showed that OCR4all can considerably outperform the commercial state-of-the-art tool ABBYY Finereader on moderate layouts if suitably pretrained mixed OCR models are available. Further on, extensive configuration capabilities are provided to set the degree of automation of the workflow and to make adaptations to the carefully selected default parameters for specific printings, if necessary. In the long run, this constant manual correction produces large quantities of valuable, high quality training material, which can be used to improve fully automatic approaches. To deal with this issue in the short run, OCR4all offers a comfortable GUI that allows error corrections not only in the final output, but already in early stages to minimize error propagations. This is mostly due to the fact that the required ground truth for training stronger mixed models (for segmentation, as well as text recognition) is not available, yet, neither in the desired quantity nor quality.
While a variety of materials can already be processed fully automatically, books with more complex layouts require manual intervention by the users. In this paper, we present an open-source OCR software called OCR4all, which combines state-of-the-art OCR components and continuous model training into a comprehensive workflow. The drawback of these tools often is their limited applicability by non-technical users like humanist scholars and in particular the combined use of several tools in a workflow. Nevertheless, in the last few years, great progress has been made in the area of historical OCR, resulting in several powerful open-source tools for preprocessing, layout analysis and segmentation, character recognition, and post-processing. Optical Character Recognition (OCR) on historical printings is a challenging task mainly due to the complexity of the layout and the highly variant typography.