Browse Prior Art Database

Method and System to Avoid Spam for Social Networks Disclosure Number: IPCOM000249586D
Publication Date: 2017-Mar-07
Document File: 5 page(s) / 137K

Publishing Venue

The Prior Art Database


Disclosed are a method and system to use image and video recognition to detect and block inappropriate content generated and posted to social media websites by individuals and groups. The system also assigns ratings for the trustworthiness of each user.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.


Method and System to Avoid Spam for Social Networks

Social networks users frequently receive invitations (invites) from people that want to follow those users. In some instances, the requests are spam and come from users that want to share unwanted or explicit content (video, images, and text).

Social networks currently have some methods to identify spam, such as a user following a large number of users in a short period, a large number of people blocking a specific user, a large number of spam complaints filed against a user , etc. Spam is constantly evolving, so these methods do not detect all types of spam, especially those that come together with explicit contents. Some methods are based on user interaction, which is an ineffective approach.

In addition, social networks can contain multiple profiles that present fake promotions (usually offering free products) in order to receive a high number of likes and shares within a short period. These sometimes share malicious code.

Figure 1: Scenario without proposed solution


The novel solution is a method and system to identify and block spam generated by individuals and groups on social networks by performing image and video recognition on posts. The system also assigns ratings for the trustworthiness of each user. The core novel methods include:

• Dynamically rating user trustworthiness based on user actions

• Using text, image, and video recognition technologies to detect inappropriate content posted in social networks

• Using image, video, and text recognition technology to analyze social network groups and identify groups behavior

By performing text, image, and video recognition, the system detects and blocks inappropriate content before forwarding it to users, instead of relying on users to see the posts and then make a decision to denounce that post . Current anti-spam methods consider an X number of users denouncing a post before blocking it , which means that some users see the post before a system removes it.

By rating user trustworthiness, the novel method avoids blocking contents from a conscientious user that accidentally shares


inappropriate content. The system notifies that user about the content that was blocked , but allows the user to continue posting to the social network.

The solution can also analyze a group’s behavior. The system automatically checks groups that receive a high number of 'likes' in a short period. It analyzes the posts using image and text recognit...