Digital Data Stream Segmentation and Signature Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for extracting and annotating sequences and inter-sequences from digital data streams, such as television streams, are imprecise, incomplete, and unable to provide on-the-fly processing, relying on manual reference sets or periodic detection methods that are not efficient for real-time analysis.
Innovation Solution
A method that splits the digital data stream into segments, calculates signatures for these segments, and performs on-the-fly inter-sequence detection using a combination of separation detection and signature comparison with an automatically updated reference base, allowing for real-time identification and annotation of sequences and inter-sequences without prior knowledge of the stream.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If programming information from EITs or program guides is used for extraction, then the extraction process is simple and straightforward, but the precision and completeness of sequence detection deteriorates
Solution Approach 1:
The patent replaces manual reference set creation and periodic global analysis with automatic on-the-fly content analysis using computational algorithms. The system automatically extracts visual descriptors from video frames, audio descriptors from audio signals, and metadata from the stream, eliminating the need for manual reference constitution while achieving real-time processing.
Solution Approach 2:
The system performs self-service by automatically creating and updating its own reference base during stream processing. Instead of relying on pre-created manual reference sets or external program guides, the system autonomously learns from the stream content itself, continuously refining its ability to detect sequences and inter-sequences without human intervention.
2Measurement precision
If a manual reference set is created for content analysis, then the initial detection accuracy is improved, but the system cannot adapt to new programs and the processing cannot be done on-the-fly
Solution Approach 1:
The reference base transitions from a static manual creation process to a dynamic automatic update process. The system continuously learns from incoming stream content, automatically adding new programs and inter-sequences to the reference base as they are detected. This dynamic adaptation allows the system to handle new content types and formats without manual intervention while maintaining detection accuracy.
Solution Approach 2:
The system performs preliminary content analysis by extracting visual, audio, and metadata descriptors from incoming stream segments before making detection decisions. This preliminary processing enables the system to quickly identify potential sequences and inter-sequences, building the reference base proactively as data flows through the system rather than waiting for periodic updates.
3Reliability
If periodic global analysis is performed for annotation, then comprehensive detection of repeated inter-programs is achieved, but real-time on-the-fly processing is not possible
Solution Approach 1:
The system implements continuous on-the-fly processing by performing content analysis, descriptor extraction, and sequence detection continuously as data flows through the system. Instead of periodic batch processing, the analysis operates continuously without interruption, enabling real-time detection and annotation of sequences and inter-sequences while maintaining comprehensive coverage through sustained monitoring.
Solution Approach 2:
The stream is divided into segments with extracted visual, audio, and metadata descriptors processed independently and continuously. This segmentation enables parallel processing of multiple stream portions simultaneously, maintaining real-time processing capability while ensuring comprehensive detection through systematic coverage of all segments as they pass through the analysis pipeline.
4Reliability
If repetitive nature of inter-sequences is used for detection, then detection of repeated inter-programs is improved, but the detection period must be long and on-the-fly processing is not enabled
Solution Approach 1:
The patent replaces periodic mechanical comparison of accumulated descriptors with continuous on-the-fly content analysis. By extracting and analyzing visual, audio, and metadata descriptors in real-time as segments pass through the system, the detection mechanism achieves high reliability without requiring long accumulation periods, enabling immediate detection of inter-sequences as they occur in the stream.
Data Source
Figure 1
Figure 2
AI summary
The invention relates to a method for processing a digital data stream including sequences and intersequences, wherein said method comprises a step (E1) of cutting the stream into segments on the fly and calculating signatures for said segments, and a step (E2) of detecting intersequences during which the following operations are applied to said stream on the fly: a first detection of intersequences by separation detection; the supply of a reference base (BR) with at least the signatures of the intersequences detected by the first detection; a second detection of intersequences by comparing the signatures of the stream segments obtained during the cutting step (E1) with the signatures of the intersequence segments contained in the reference base (BR).