Abramowitz and Stegun - A Resource for Mathematical Document Analysis

摘要

In spite of advances in the state of the art of analysis of mathematical and scientific documents, the field is significantly hampered by the lack of large open and copyright free resources for research on and cross evaluation of different algorithms, tools and systems. To address this deficiency, we have produced a new, high quality scan of Abramowitz and Stegun's Handbook of Mathematical Functions and made it available on our web site. This text is fully copyright free and hence publicly and freely available for all purposes, including document analysis research. Its history and the respect in which scientists have held the book make it an authoritative source for many types of mathematical expressions, diagrams and tables. The difficulty of building an initial working document analysis system is a significant barrier to entry to this research field. To reduce that barrier, we have added intermediate results of such a system to the web site, so that research groups can proceed on research challenges of interest to them without having to implement the full tool chain themselves. These intermediate results include the full collection of connected components, with location information, from the text, a set of geometric moments and invariants for each connected component, and segmented images for all plots.