Browse Prior Art Database

A method of recommending and inserting graphics by using Cognitive API Disclosure Number: IPCOM000248932D
Publication Date: 2017-Jan-22
Document File: 3 page(s) / 63K

Publishing Venue

The Prior Art Database


This disclosure provides a method of recommending and inserting graphics in pages such as slides intelligently. The method recommends graphics for the context based on cognitive technology.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 52% of the total text.


A method of recommending and inserting graphics by using Cognitive API

To recommend suitable graphics in pages, the following steps are required:

1. Analyze the contents of the working page, extract the keywords based on the contents a. Analyzing the contents, extracting the words existing in the page b. Rating the words according to the existence times c. Defining them as the input that will be sent to the background analyzer server

2. Analyze server proceed searching and figure out the proper candidates images a. Analyze server proceed searching according to the input from step 1. b. The analyze server will rate the images by different sequence such like: size\popularity\clearance and other conditions that could be defined by the users c. The analyze then sends the first 10 images with top rate back to the clients. Upon client's request, the server will provide other options of the images

3. Push recommendations to users.


Step 1: Sending the content of the previous page (along with other pages, other pages are used as context) to Cognitive server via API; Input all the content of the previous page, instead of providing keyword or high rating word to the Cognitive server. Via this way, the meaning of the page is calculated more accurately by the Cognitive, based on the background of the previous page.

Step 2: Sending the abstract info out of the first step to Cognitive server via API; Call the API to get the images which could help interpret the topic on the previous page; The images will be grouped and prioritized by the Cognitive server. Use the output of Step 1 as input, then call the Cognitive API to complete the search. It will figure...