Sequential Multimedia Filtering via Accumulated Probability Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for filtering malicious multimedia services are ineffective due to reliance on simple features and delayed analysis, leading to low classification performance and exposure of users to harmful content, especially in real-time services.
Innovation Solution
A malicious multimedia filtering apparatus and method using sequential data processing, which trains a classification model on features extracted from multimedia data and determines maliciousness by calculating an accumulated probability ratio over time, allowing for real-time classification and filtering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the entire data of a multimedia service is received and analyzed to determine maliciousness, then the classification accuracy is improved, but the storage space requirement and analysis time increase significantly
Solution Approach 1:
The patent divides multimedia data into sequential segments and processes them in order, analyzing only the necessary portion of data to make a classification decision. This segmentation approach allows the system to achieve accurate classification without requiring to store and analyze the entire multimedia dataset, thus resolving the contradiction between classification accuracy and analysis time.
Solution Approach 2:
The patent extracts and analyzes key features from multimedia data segments in advance during sequential processing, rather than waiting to receive and process all data. This preliminary feature extraction and analysis enables the system to make classification decisions based on sufficient information without requiring complete data reception, thereby reducing both storage requirements and analysis time while maintaining classification accuracy.
2Productivity
If simple features are used for maliciousness determination, then the processing speed is improved, but the classification performance deteriorates
Solution Approach 1:
The patent applies different feature extraction and analysis methods to different segments of multimedia data based on their local characteristics. By adapting the analysis depth and feature types to specific data segments, the system achieves high classification performance without uniformly applying complex processing to all data, thus maintaining high processing speed while improving classification accuracy.
Solution Approach 2:
The patent dynamically adjusts the feature extraction and analysis process during sequential data processing based on the accumulated information and confidence levels. As more data segments are processed, the system can transition from simple to more complex feature analysis only when necessary, optimizing the balance between processing speed and classification performance adaptively.
3Speed
If data is received and processed one by one in real time, then the response time is improved, but the classification accuracy deteriorates due to limited data availability
Solution Approach 1:
The patent implements a feedback mechanism in the sequential processing system where classification decisions are continuously refined as more data segments are received. The system uses feedback from previous segment analyses to adjust subsequent processing strategies, allowing it to maintain high response time by making preliminary classifications while improving accuracy through iterative refinement as additional data becomes available.
Solution Approach 2:
The patent performs preliminary classification based on features extracted from early data segments, enabling real-time response. This preliminary classification is then refined as more data segments are processed sequentially, allowing the system to provide immediate responses while continuously improving classification accuracy without delaying the initial response time.
Data Source
AI summary
An apparatus for filtering malicious multimedia data using sequential processing and a method thereof are provided. The apparatus includes: a maliciousness classification model training unit extracting a predetermined feature from at least one or more types of moving pictures and then, through machine training, generating a maliciousness determination model for each of at least one or more classes; a malicious data classification unit sequentially inputting input moving pictures for which maliciousness is required to be determined, to the maliciousness determination model, and determining the maliciousness class of the input moving pictures, based on a probability that data at a determination time of the input moving pictures belongs to a predetermined maliciousness class, and an accumulated maliciousness probability to a current time; and a malicious information filtering unit cutting off service if the maliciousness class belongs to a predetermined reference maliciousness class.