Browse Prior Art Database

Method to Establish Relationships Using Visual Cognitive Insights

IP.com Disclosure Number: IPCOM000248995D
Publication Date: 2017-Jan-25
Document File: 5 page(s) / 83K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a cognitive system that develops insights and confidence levels around how well two people appearing in a digital picture know each other. The system utilizes cognitive classifiers, scoring, insights, and confidence levels. The system does not require the use of tagging or visual identification of the subjects selected.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 33% of the total text.

1

Method to Establish Relationships Using Visual Cognitive Insights

Social networks provide a multitude of data as to how people are connected. If a system or application can recognize the social network identities of two people, then it can use a wide variety of data sources to determine the relationship and between those two people and the depth of that relationship. Existing methods ask questions such as:

 Are these two people connected on a social network?  Do these people have common friends on social networks?  Do these people follow similar interest groups on social media?  Are these people tagged together in photographs?

Existing art can establish the relationship of people in untagged photographs. This can determine, for example, if an untagged person is a relative of other people shown in a picture. Other methods can establish relationships in non-visual ways.

No current art enables a user to select two people in a photograph and generate insights that describe how well these people know each other and in what capacity.

The novel contribution is a cognitive system that develops insights and confidence levels around how well two people appearing in a digital picture know each other. The system utilizes cognitive classifiers, scoring, insights, and confidence levels. The system does not require the use of tagging or visual identification of the subjects selected.

The novel method determines the relationship of two people appearing in a picture where the identities of those people are unknown. These people are not tagged in the picture and the respective social networks are not known; therefore, no current method can analyze the photograph to determine a pre-existing relationship. This method can be applied to any digital picture of multiple people to answer questions such as:

 How well do these people know each other?  What type of relationship do these people have?

During implementation of the novel system, a user selects two people in a digital photograph that is posted on social media. The cognitive system analyzes the photograph to generate classifiers. A cognitive system scores the classifiers to indicate the classifiers’ relevance in establishing a relationship between two people. The system then generates cognitive insights with confidence levels. In addition, the system provides the detection and analysis of other digital photographs to provide cognitive insights and confidence levels.

The method to establish relationships using visual cognitive insights comprises five stages.

Stage 1: User selects two people in a digital photograph posting A user opens a social network post, such as a photograph posted on social media, featuring a picture that shows two or more people. The user selects two people in the

2

picture. These two people are not tagged, meaning the identities and social network profiles are not known. Alternatively, the system can proactively select the two most well defined faces and automatically begin to an...