Video Pushing System with Photosensitive Recognition Model

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Many video creators lack knowledge about photosensitive epilepsy, leading to the potential triggering of photosensitive epilepsy in viewers, which can cause discomfort or health issues in sensitive individuals.

Innovation Solution

A method and apparatus for video transmission that utilize a photosensitive video recognition model to tag videos as either photosensitive or not, allowing users to filter out photosensitive videos and improving video viewing safety.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If photosensitive videos are pushed to browsing users, then video content diversity and user engagement are improved, but health safety and user comfort deteriorate due to potential photosensitive epilepsy triggering

Engineering Contradiction:
Improvevideo content delivery efficiencyVSAvoidphotosensitive epilepsy risk
Core Design Contradiction:
ProductivityVSObject-affected harmful factors

Solution Approach 1:

The system performs preliminary classification of videos into photosensitive and non-photosensitive categories before pushing to users. A recognition model pre-processes video content to identify photosensitive characteristics, and filtering logic is pre-established to prevent harmful videos from reaching vulnerable users, thus eliminating the harm before it occurs.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary filtering system between video content and users. This intermediary includes a recognition model that classifies videos and a pushing logic that selectively delivers content based on user sensitivity profiles, acting as a mediator that protects users while maintaining content delivery functionality.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Reliability

If a photosensitive video recognition model is implemented to filter videos, then user safety is improved, but system complexity and processing time increase

Engineering Contradiction:
Improvevideo viewing safetyVSAvoidsystem structure complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The system segments the video processing function into distinct modules: video acquisition, recognition model processing, classification logic, and selective pushing. This segmentation allows each component to be independently optimized and maintained, reducing overall system complexity while achieving reliable photosensitive video filtering.

Inventive Principle:
Principle #1Segmentation

3Object-affected harmful factors

If photosensitive videos are filtered out from video pushes, then user comfort is improved, but video content variety and user interest matching deteriorate

Engineering Contradiction:
Improvephotosensitive video impactVSAvoidvideo content adaptability
Core Design Contradiction:
Object-affected harmful factorsVSAdaptability or versatility

Solution Approach 1:

The system applies local quality filtering by differentiating between photosensitive and non-photosensitive video content, and between sensitive and non-sensitive users. The filtering is selectively applied only where necessary (to photosensitive videos for sensitive users) while preserving full content access for others, thus maintaining content variety and adaptability without compromising user comfort.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4550801A1Method and apparatus for pushing video, and storage medium
Publication Date: 2025.05.07 DOUYIN VISION CO LTD
  • EP4550801A1 patent drawingFigure 1~2
  • EP4550801A1 patent drawingFigure 3~4
  • EP4550801A1 patent drawingFigure 5~6

AI summary

The present disclosure relates to the technical field of multimedia, and in particular to a method and apparatus for pushing a video, and a storage medium. The method for pushing a video comprises : acquiring one or more videos, the one or more videos comprise a video having a photosensitive identifier and a video having no photosensitive identifier, wherein each video has a photosensitive identifier if the video is determined to be a photosensitive video by a photosensitive video recognition model; and if a browsing user is in a state of filtering the photosensitive video, pushing, to the browsing user, the video having no photosensitive identifier in the one or more videos.