Trainable separation

Trainable separation classifies each individual page and then based on those classification results, separates the pages into documents. Trainable separation is useful when the length of a document is not consistent.

Trainable separation is disabled by default.

Separation and classification are performed at the same time during production and the same documents are used to train both classification and separation. During design-time, it is not possible to test trainable separation. Only classification is testable at design-time.