Classification - ABBYY Documentation

ABBYY FlexiCapture 12 features technologies that let you classify documents and detect their types and variants without creating FlexiLayouts in FlexiLayout Studio. Classification can be used to sort documents, detect specific documents among all input documents, and assemble input pages into documents. Classification scenarios Before a classifier can be used to classify documents it must be trained. To train a classifier, you need to load sample documents and assign appropriate classes to these documents manually. Once trained, the classifier will then be able to classify new documents on its own. Setting up a classifier ABBYY FlexiCapture 12 includes trainable technologies for classifying documents based on their images (for documents that can be differentiated visually) and on their text. These technologies can be automatically combined, in which case the program will choose the most suitable technology based on the nature of the analyzed document. Classification profiles Some classification scenarios may favor precision over recall, while in others recall may be more important. The program allows you to prioritize precision over recall, recall over precision, or strike a balance between the two. Types of classification errors and balance between precision and recall As an additional classification aid, developers can use ABBYY FlexiLayout Studio, which can handle thousands of images and offers tools for thorough analysis of classification results and fine-tuning classifiers, including rules for handling text and visual objects on pages. Importing a classifier from ABBYY FlexiLayout Studio Classification is the process of putting document images or text documents into certain predefined categories. Documents can be classified manually or automatically. Manual classification of large amounts of documents is

labor-intensive
costly
slow
error-prone when working against tight deadlines

ABBYY FlexiCapture 12 includes classification technologies that can be easily trained to classify documents without creating complicated rules or FlexiLayouts in ABBYY FlexiLayout Studio. The ABBYY FlexiCapture classification technologies are highly flexible and can be used to distinguish:

Document types (e.g. you can separate passports from invoices)
Document variations (e.g. you can detect invoices received from different vendors or driver’s licenses issued in different states)
Document boundaries (e.g. you can detect the first and the subsequent pages of documents of each type).

​See also:

See also: