US 12,639,377 B2 · Resolving points of interest by text stylization

Abstract

A screen reader application traverses each node in a document object model (DOM) for the text stylization. Properties for foreground color, background color, font type, font size and font stylization are algorithmically reduced to an identifier. Each node in the DOM with the same identifier has the same text stylization. Unique and infrequent text stylizations by a webpage author signal a point of interest. The screen reader application locates and navigates to that node in the DOM on behalf or in response to the end user. Points of interest are further identified by a number of additional factors. A first includes percentage of text of having the text stylization versus total text in the DOM. A second includes excluding candidate point of interest nodes having more than 250 characters. Others include imposing minimum font sizes and text contrast ratios to qualify as a point of interest.

Claims20

Dependent claims

A method comprising:

accessing, over the Internet and by a processor, a document object model (DOM) of a web page, the DOM having a plurality of nodes;
automatically traversing each node in the DOM, by the processor, for a collection of at least two text attribute properties of each node in the DOM, the text attribute properties in the collection selected from the group consisting of foreground color, background color, font type, font size and font stylization;
reducing each collection of text attribute properties by the processor, to a style identifier whereby nodes in the DOM with identical text attribute properties have the same style identifier;
determining an atypical style identifier, by the processor, based on its frequency of appearance in the nodes in the DOM;
and
marking, by the processor, all the nodes in the DOM matching the atypical style identifier whereby a screen reader application is operable to navigate directly to those matching nodes in the DOM wherein the screen reader outputs Braille or audio.

3 more elements

A method comprising:

accessing, over the internet and by a processor, a document object model (DOM) of a web page, the DOM having a plurality of nodes;
automatically traversing each node in the DOM, by the processor, for a collection of at least two text attribute properties of each node in the DOM, the text attribute properties in the collection selected from the group consisting of foreground color, background color, font type, font size and font stylization;
reducing each collection of text attribute properties, by the processor, to a style identifier whereby nodes in the DOM with identical text attribute properties have the same style identifier;
calculating, by the processor, a point of interest value for each style identifier wherein the point of interest value comprises a relative uniqueness of each style identifier within the document as a whole;
and
comparing, by the processor, the point of interest value with a threshold uniqueness value wherein if the point of interest value exceeds the threshold uniqueness value then a point of interest text string in the DOM having the same style identifier is marked whereby a screen reader application is operable to navigate directly to the point of interest text string in the DOM and the screen reader broadcasts audio the point of interest text string.

3 more elements

Dependencies are inferred from the claim text.

Resolving points of interest by text stylization

Life of this patent

Resolving points of interest by text stylization

Life of this patent

Abstract

Claims20