SignatureDetectDOCLIB Documentation

Current Version: SignatureDetectDOCLIB v1.0

DLSignatureDetect implements a multi-scale signature detection and segmentation approach. The detection algorithm works in one of the two modes specified by the user. It can either detect signatures by looking at the entire document image, or detect signatures by exploring document context. When using document context, it effectively estimate statistics of machine printed text lines and use them to locate the region below the main body of the document text, where signatures typically appear. Using document context is more effective for machine printed documents.

The idea of this signature detection approach is to capture the structural saliency of a signature by measuring its dynamic curvature without recovering the tempo information. Once the most salient part of the signature is identified, contour grouping is performed to obtain a complete and segmented signature. As tested in large real world datasets, this approach is robust under large intra-class variations that typically exhibited on unconstrained handwritting, and it is very effective across language differences.


SignatureDetectDOCLIB Library is an add-on of DOCLIB. DOCLIB is being developed under contract by a collaboration between:
The Laboratory for Language and Media Processing
Unviersity of Maryland, College Park
and
Booz | Allen | Hamilton

All Rights Reserved, 2003-2007