Computational method for document object locator combination

摘要

This paper presents a method for combining multiple document object locators tuned to different object characteristics, with the goal of achieving location performance excelling that of any individual locator. The method includes (ⅰ) a scheme for consistent representation of locator outputs regardless of output levels, (ⅱ) the notion of object correspondence and their applications to determining what decisions to combine, (ⅲ) a mechanism for representing knowledge of locators and its use for dynamic locator selection, and (ⅳ) functions for combining confidence values of objects. Results from experiments in postal address block location using three locators and 1,100 envelope images are presented.