Browse Prior Art Database

Automated Advanced Figure Queries from GUIs in Technical Papers

IP.com Disclosure Number: IPCOM000247533D
Publication Date: 2016-Sep-14
Document File: 3 page(s) / 39K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed are a method and system that enable a user to perform automated advanced figure queries from graphical user interfaces (GUIs) in technical papers. The system uses automatic graphical input from a first technical figure, based on a GUI trigger request, and automatic semantic input related to that figure to perform a search and return related figures or produce novel or composite figures.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 36% of the total text.

Page 01 of 3

Automated Advanced Figure Queries from GUIs in Technical Papers

A method is needed to assist researchers (e.g., healthcare professionals or researchers examining trends) in locating relevant figures or related information within technical papers.

The novel contribution is a method and system that enable a user to perform automated advanced figure queries from graphical user interfaces (GUIs) in technical papers. The system uses automatic graphical input from a first technical figure, based on a GUI trigger request, and automatic semantic input related to that figure to perform a search and return related figures or produce novel or composite figures.

The GUI trigger is supplied by one or more GUI elements associated with the first figure (e.g., buttons or sliders at the edge of a figure in the portable document format (PDF) of a technical figure). The system uses the first figure to query sensors, information repositories, etc., and return desired information on discovered correlations.

The semantic input comes from a latent semantic indexing (LSI) of the figure caption or the text of the entire paper. Semantic input can be, but is not limited to, textual labels/numbers on a graph (e.g., axes labels, keys, etc.), the text of a figure caption, tags and metadata, the text in the paper in proximity to a figure that reference/describes it, text in an abstract, etc. Related figures can be, but are not limited to, figures with similar mathematical and/or complementary relationships, trends, figures with the same or similar axes legends, etc. The system sorts any returned figures by any of the following: degree of relevance, journal name, field of journal, accessibility of a journal for a given user (e.g., open source or behind a pay wall), a specified willingness of authors to share data, the presence of error bars, author information, paper citation information, etc. If the system does not discover a similar relationship, then it reports that to the user.

The system incorporates the information from the related figures into the first technical figure (e.g. by using the supplied GUI interface). Alternatively, it cancreate new composite figures from data mined from tables and text across one or more papers and other figures. The system can store newly created figures and subsequently use those for later searches.

Technical journals may embed a standardized special GUI (e.g., in a PDF or similar document or online version) at the side of figures to allow all users (or only special paying users, or journal subscribers) to specify aspects of a figure query. Alternatively, a user may "right click on a figure" to pop up a menu of features that is useful for specifying aspects of a search related to the figure or figures.

The system can use the GUI element for constructing a hypothetical figure that represents the axes, scale, and trend for which the user may wish to search in the literature. The axes may be labeled with units (e.g., meters/...