Browse Prior Art Database

A solution to cross-refer chart and specifications in document Disclosure Number: IPCOM000194428D
Publication Date: 2010-Mar-24
Document File: 6 page(s) / 114K

Publishing Venue

The Prior Art Database


We often see some complicated graph with many icons in it, like the FIG. 4 in Figure 1. The icons’ specifications usually appeared in many pages. It’s very inconvenient to read and refer among figure and its specifications. For example, in Figure 1, FIG. 4 is a graph to specify a work flow. There is an icon named 401 pointed to a component. In the document, there are 6 places which referred to 401 and give detailed descriptions. When we read at the 6 places, we need to refer to FIG 4, or when we look at FIG 4, we also need to refer to its specifications. In old solutions, we have to jump (switch) among document’s pages to find the graph or specifications. We usually use ‘Searching’ or ‘Page Down/Up’ actions to find the places by ourselves. Although some searching methods can list all results to let us choosing, we still just can go to one paragraph at one time. There isn’t effective approach of building bi-directional connections between graphs and their specifications, nor convenient layout for users browsing them.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 6

A solution to cross

A solution to crossA solution to cross -

---refer chart and specifications in documentrefer chart and specifications in documentrefer chart and specifications in document

refer chart and specifications in document



Our approach to resolve such a problem is to build up bi-directional connections automatically between graphs and their specifications, and, provide a convenient layout for user to read graphs and specifications.

An example to specify our solution is like Figure 1. Icon 401 in FIG.4 have 6 specifications in the document. By adopting our solution, when user highlight icon 401 (e.g. by hover on it), all the 6 referred places will be listed in another display layer ( Search result list layer) , including their abstracts and page info. When user chose one abstract, detail information will be shown on the third display layer Detail of one result layer . We can design these two additional display layers as half-transparent / fade-in / fade -out. We can lay out these layers out of original document region too.

Figure 1


Display of graph and icons' references in our solution



To implement the solution, our method mainly contains 5 steps,

1) Abstract text in graphs by automatically image-to-text method.
2) Use abstracted icon texts as key words to search in the whole document, extract search result abstracts and position information. (Existent effective document abstracting algorithms can help us to implement this step).

3) Build up bi-directional mapping between icons texts and document paragraphs.
a) Create two table to store bi-directional mapping relationship Text-


Details table

: Indexing from icon text to search results (sorted by


[This page contains 1 picture or other non-text object]

Page 2 of 6

relevance), and to details of every search result
Word-Text-Figure table : Indexing from words in document paragraphs to icon text, and to figure
b) For each icon text, algorithm also abstract words in figure's surrounding paragraphs to assistant exactly identify figure's specifications, and to impact search results' relevance and sorting. (Existent figure surrounding algorithms can help us to implement this step)

c) Store bi-directional mapping index to above tables
4) Add user interaction interface to figures and words
5) When user hovering on a graph icon, search result list layer which contained all the abstracts will be shown besides the icon. When user choosing one abstract, detail of one result layer will be displayed.


))System Overview

In Figure 2, it specified the 4 components in our method,

S1: User Interaction component

, which deal with user actions like 'hover', 'choose', etc.

                   , which analysis figures and texts to build up the bi-directional index, and add user response interface for S1.


Preprocess component

S3: Visualization component

, which response to user actions, and create display layer based on reference



Indexing comp...