Report Reliability Evaluation for Safer Random Video Calls
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users making video calls are exposed to inappropriate content, leading to feelings of sexual shame or offense, which existing technologies fail to adequately address.
Innovation Solution
A report evaluation device and method that utilizes learning models to categorize and evaluate the reliability of reports on inappropriate content, employing algorithms to censor video, audio, text, and gestures, and manage video call sessions to prevent exposure to such content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If video call services are provided to enable random video chat between users, then communication convenience and user engagement are improved, but users may be exposed to inappropriate content causing sexual shame or offense
Solution Approach 1:
The system performs preliminary evaluation of user reports before inappropriate content reaches other users. When a user reports inappropriate content, the server proactively evaluates the report using learning models and takes preventive action by notifying other users involved in the video call, thereby preventing potential exposure to harmful content before it occurs.
Solution Approach 2:
The server acts as an intermediary between users during video calls. It receives reports from users, evaluates them using learning models, and mediates the situation by notifying relevant users of potential inappropriate content. This intermediary function allows the system to maintain video call accessibility while protecting users from harmful content through automated moderation.
2Measurement precision
If manual report evaluation is used to filter inappropriate content, then content accuracy can be maintained, but system complexity and response time increase
Solution Approach 1:
The system uses self-service through automated learning models that independently evaluate user reports without requiring manual human review. The server receives reports, automatically classifies them using trained learning models (such as image recognition or text analysis models), and determines whether content is inappropriate. This self-service approach maintains high evaluation accuracy while reducing system complexity and enabling faster response times compared to purely manual evaluation.
Solution Approach 2:
The patent replaces manual mechanical evaluation processes with automated electronic learning models. Instead of relying on human reviewers to assess each report, the system uses machine learning algorithms (image recognition, text analysis) to automatically evaluate content. This substitution reduces system complexity by eliminating manual intervention while maintaining or improving evaluation accuracy through consistent automated assessment.
3Reliability
If comprehensive learning models are deployed to evaluate all report types, then content safety is improved, but processing time and computational resources increase
Solution Approach 1:
The system segments the evaluation process by categorizing reports into different types (image content, text content, audio content) and applying specialized learning models appropriate for each type. Instead of using a single comprehensive model for all reports, the server divides evaluation tasks into segments matching the content type, which improves both safety accuracy and processing efficiency by using the most suitable evaluation method for each specific case.
Solution Approach 2:
The system applies partial action by selectively evaluating only the necessary portions of reports based on their type. For example, image reports are evaluated using image recognition models while text reports use text analysis models, avoiding the computational overhead of applying all possible evaluation methods to every report. This selective approach maintains high content safety while reducing processing time and computational resource consumption.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A report evaluation method of a report evaluation device may include receiving a report from at least one client terminal, determining a category of the received report, identifying a learning model corresponding to the category, evaluating a reliability of the report through the learning model, and generating and outputting information on the reliability.