System and method for binary persistence format for a recognition result lattice

摘要

A system and method for storing document data in a serialized binary format recognition lattice structure so that the data is accessible to other applications. The lattice structure is generated that includes root node data. Child nodes correspond to columns of the recognition lattice. Each node includes a descriptor that is a collection of flags indicating presence of specific components of the node. The child nodes may include grandchild nodes with similar structure corresponding to elements within the columns. Each node further includes property information associated with the node. The recognition lattice is stored in a serial binary data format.