Video Stream Commercial Detection via Segment Fingerprinting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for identifying commercials in video data streams are ineffective due to limitations in accurately detecting commercial locations, and there is a need for improved approaches to identify both commercials and repeating content.
Innovation Solution
The method involves receiving and parsing video data streams into segments, creating segment representations (fingerprints), storing these in a database, and identifying matching segments to detect potential commercials and repeating content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional feature-based detection or recognition-based detection methods are used to identify commercials, then the detection process can be performed, but the accuracy of detecting commercial locations is insufficient and repeating content cannot be reliably identified
Solution Approach 1:
The video data stream is divided into multiple segments, and each segment is parsed into individual frames. This segmentation allows the system to analyze specific portions of the video independently, creating frame representations that can be compared against a database to accurately identify commercial content and repeating patterns without requiring the entire video to be processed as a single unit.
Solution Approach 2:
The system creates representations (copies) of video frames by extracting key visual features and generating hash values. These frame representations serve as digital fingerprints that can be stored in a database and compared against future frames. This copying approach enables accurate identification of repeating content and commercial locations without requiring the original video data to be retained, improving both precision and efficiency.
2Measurement precision
If the entire video data stream is analyzed to identify commercials, then complete coverage is achieved, but the processing time and computational resources increase significantly
Solution Approach 1:
The video stream is parsed into discrete segments and frames, allowing the system to process and analyze only relevant portions rather than the entire continuous stream. This enables efficient identification of commercial content while maintaining complete coverage through systematic segment-by-segment analysis.
Solution Approach 2:
The system performs preliminary processing by extracting frame representations and generating hash values in advance, storing them in a database for quick comparison. This preliminary action prepares the data structure ahead of time, so when actual commercial detection is needed, the system can quickly compare incoming frames against the pre-prepared database without performing heavy computation in real-time, significantly reducing processing time.
3Reliability
If frame representations are stored in a database for comparison, then repeating content can be identified, but the storage requirements and database complexity increase
Solution Approach 1:
Instead of storing complete video frames or large amounts of visual data, the system creates compressed representations of frames by extracting key features and generating hash values. These compact representations serve as efficient proxies that occupy minimal storage space while retaining the essential information needed for identifying repeating content and commercials through database comparison.
Data Source
AI summary
Methods and apparatus are provided for detecting potential repeating content, such as commercials, in a video data stream by receiving one or more video data streams, parsing each video data stream into a plurality of segments, creating a representation of each segment such as a segment fingerprint, storing the plurality of segment representations in a database, and identifying any segment representations in the received and parsed one or more video data streams that match segment representations in the database that were previously stored from video data streams that were previously received and parsed.