An optimization approach for invoice image analysis

Automated invoice processing and information extraction has attracted remarkable interest from business and academic circles. Invoice processing is a very critical and costly operation for participation banks because credit authorization process must be linked with real trade activity via invoices. Therefore this paper proposes a new invoice parsing method that uses a two-phase optimization structure and eliminates invoice classes. The first phase uses individual invoice part detectors such as SVM, maximum entropy and HOG to produce candidates for the various types of invoice parts. At the second phase, the basic idea is to parse an invoice by parts arranged in a deformable composition. This system can handle any type of invoice because PBM is an optimization based method. The proposed system is tested with real invoices and found to be promising for real world employment.