Video Classification Module for Detecting Looping Content

Solution Overview

Problem

Conventional content ranking systems in social networking platforms are vulnerable to manipulation, as users attempt to exploit preferences for video content over static images or live streaming by converting static images or text-only posts into videos, and these systems lack effective tools to distinguish between diverse content types at granular levels, leading to user fatigue from repetitive viral content.

Innovation Solution

The implementation of a video classification module that identifies and categorizes videos into types such as static image, slideshow, looping, poll, and meme videos by analyzing interest points, frame comparisons, downsampling, and machine learning models to filter out dynamic regions and synthetic text, allowing for accurate content classification and ranking.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional content ranking systems prefer video content over static images, then user engagement with video content increases, but users can manipulate the system by converting static images into videos, reducing content diversity and authenticity

Engineering Contradiction:

Improveuser engagementVSAvoidcontent authenticity

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent segments video content into individual frames and analyzes them separately to detect whether a video is actually a static image disguised as video. By examining frame-by-frame characteristics, the system can identify manipulated content while still allowing genuine video content to receive preferential treatment, thus maintaining both user engagement and content authenticity

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary classification system that acts as a mediator between video content and the ranking algorithm. This intermediary layer classifies video content into genuine videos versus static images disguised as videos, allowing the ranking system to treat them differently and prevent manipulation while preserving engagement with authentic video content

Inventive Principle:

Principle #24Intermediary (Mediator)

2Device complexity

If the system lacks effective tools to distinguish between diverse content types, then content classification is simple, but users experience fatigue from repetitive viral content and manipulation

Engineering Contradiction:

Improveclassification system complexityVSAvoiduser fatigue from repetitive content

Core Design Contradiction:

Device complexityVSObject-generated harmful factors

Solution Approach 1:

The patent segments video content into individual frames and analyzes them separately to detect whether a video is actually a static image disguised as video. By examining frame-by-frame characteristics, the system can identify manipulated content while still allowing genuine video content to receive preferential treatment, thus maintaining both user engagement and content authenticity

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different analysis methods to different portions of video content. It examines specific frames, detects static regions versus dynamic regions, and identifies characteristic patterns of manipulated content. This localized quality assessment allows the system to distinguish content types effectively without requiring complex global analysis of entire video sequences

Inventive Principle:

Principle #3Local quality

3Reliability

If the system downranks static image videos to prevent manipulation, then content authenticity improves, but genuine static image content may be unfairly penalized

Engineering Contradiction:

Improvecontent authenticityVSAvoidfalse penalization of genuine content

Core Design Contradiction:

ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent employs feedback mechanisms where the classification system continuously learns from identified patterns of manipulated content. By analyzing characteristics of downranked content and comparing with genuine video content, the system refines its detection algorithms to reduce false positives, ensuring that genuine static image content is not unfairly penalized while maintaining authenticity detection

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent introduces an intermediary classification system that acts as a mediator between video content and the ranking algorithm. This intermediary layer classifies video content into genuine videos versus static images disguised as videos, allowing the ranking system to treat them differently and prevent manipulation while preserving engagement with authentic video content

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS10956746B1Systems and methods for automated video classification

Publication Date: 2021.03.23 META PLATFORMS INC

AI summary

Systems, methods, and non-transitory computer-readable media can receive a set of video frames associated with a video. A determination can be made that a first set of consecutive video frames of the set of video frames depicts identical content to a second set of consecutive video frames of the set of video frames, wherein the first set of consecutive video frames and the second set of consecutive video frames satisfy a threshold number of consecutive video frames. The video is identified as a looping video based on the determination that the first set of consecutive video frames depicts identical content to the second set of consecutive video frames.