Back to results
US 9,753,901 B1Granted
Identifying important document content using geometries of page elements
AssigneeGoogle Inc.
InventorsYifan Xu, Xiaofeng Mi
Life of this patent
Estimating the term…
Abstract
Systems and techniques are provided for detecting columns of an electronic page based on a render of the electronic page and identification of one or more columns based on the render. A column of interest may be identified based on detecting the one or more columns based on a physical position of the column, a column size, and/or the content within the column. The column of interest may be used to index or categorize the electronic page as the most relevant information corresponding to the page may be contained in the column of interest.
Claims24
A computer implemented method, comprising:
- receiving render nodes for a rendered electronic page, each render node corresponding to a respective document object model tree node, and each render node including respective boundary information corresponding to a structured element of the respective document object model tree node to which the render node corresponds, wherein the render nodes are part of a render tree;
- identifying, from the respective boundary information and based on an outlier threshold value, one or more outliers of the rendered electronic page, wherein an outlier is content that is not rendered as a user-viewable portion of the rendered electronic page, and wherein the outlier threshold value designates a boundary on the rendered electronic page where information outside of the boundary contains outlying content for the rendered electronic page, the designation based on where a highest percentage of content is located within the rendered electronic page;
- excluding, from the respective boundary information, boundary information that is outside of the outlier threshold value, including the one or more identified outliers;
- detecting, from the respective boundary information of the render nodes, a plurality of page columns of the rendered electronic page, wherein the respective boundary information excludes the boundary information of the one or more identified outliers;
- determining, from the plurality of page columns and the boundary information of the render nodes, that one of the page columns is a column of interest, the determination being based on criteria related to positions of the column in the rendered electronic page;
- and
- utilizing, by a searching engine, during classifying the rendered electronic page, the content in the column of interest associated with the rendered electronic page in a manner that is different from the utilizing of content from other columns within the rendered electronic page, wherein utilizing of content from other columns within the rendered electronic page comprises assigning a weight for content within the column of interest that is higher than weights assigned to content outside of the column of interest.
4 more elements
A system, comprising:
- a data processing apparatus including one or more hardware processors;
- and
- a memory storage apparatus in data communication with the data processing apparatus, the memory storage apparatus storing instructions executable by the data processing apparatus and that upon such execution cause the data processing apparatus to perform operations comprising:
- receiving render nodes for a rendered electronic page, each render node corresponding to a respective document object model tree node, and each render node including respective boundary information corresponding to a structured element of the respective document object model tree node to which the render node corresponds, wherein the render nodes are part of a render tree;
- identifying, from the respective boundary information and based on an outlier threshold value, one or more outliers of the rendered electronic page, wherein an outlier is content that is not rendered as a user-viewable portion of the rendered electronic page, and wherein the outlier threshold value designates a boundary on the rendered electronic page where information outside of the boundary contains outlying content for the rendered electronic page, the designation based on where a highest percentage of content is located within the rendered electronic page;
- excluding, from the respective boundary information, boundary information that is outside of the outlier threshold value, including the one or more identified outliers;
- detecting, from the respective boundary information of the render nodes, a plurality of page columns of the rendered electronic page, wherein the respective boundary information excludes the boundary information of the one or more identified outliers;
- determining, from the plurality of columns and the boundary information of the render nodes, that one of the page columns is a column of interest, the determination being based on criteria related to positions of the column in the rendered electronic page;
- and
- utilizing, by a searching engine, during classifying the rendered electronic page, the content in the column of interest associated with the rendered electronic page in a manner that is different from the utilizing of content from other columns within the rendered electronic page, wherein utilizing of content from other columns within the rendered electronic page comprises assigning a weight for content within the column of interest that is higher than weights assigned to content outside of the column of interest.
7 more elements
Dependencies are inferred from the claim text.