Video Recording Apparatus Scene Change Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video recording apparatuses face challenges in detecting scene changes and commercial message broadcasts in real-time during video recording, with existing methods being inefficient due to reliance on encoder chip sets, offline operations, inaccurate black frame detection, and difficulty in identifying CM broadcast periods with high accuracy, especially when audio schemes are similar or when CM lengths are not integer multiples of 15 seconds.
Innovation Solution
A video recording apparatus that encodes video data, decodes the stream to extract scene changes, and detects silence periods simultaneously, using a CM-broadcast-period determination mechanism to identify CM broadcast periods by successively detecting changeover-points within a predetermined time range, allowing for real-time extraction and recording of scene changes and CM periods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If scene change detection is performed using encoder chip sets, then scene change extraction can be integrated into the encoding process, but the detection accuracy is limited by encoder capabilities and requires adjustment for different encoder types
Solution Approach 1:
The patent extracts the scene change detection function from the encoder by introducing a separate detection unit that operates independently. This unit reads video data from the recording medium and performs histogram comparison to detect scene changes, rather than relying on the encoder's internal mechanisms. This separation allows for more accurate and encoder-independent detection.
Solution Approach 2:
The patent introduces a detection unit as an intermediary component between the video data storage and the scene change information generation. This intermediary unit uses histogram analysis and difference calculation to bridge the gap between raw video data and meaningful scene change detection, providing a standardized interface independent of encoder variations.
2Ease of manufacture
If scene change detection is performed on already stored video data, then detection can be done offline, but it requires additional time to read and process the stored data
Solution Approach 1:
The patent performs scene change detection during the video recording process itself, rather than as a separate post-processing step. The detection unit operates concurrently with the recording process, analyzing video data as it is being stored. This preliminary action eliminates the need for separate offline processing and reduces the time loss associated with later data retrieval and analysis.
3Measurement precision
If audio scheme-based detection is used to identify CM broadcasts, then program main parts and CM broadcasts can be distinguished, but the method fails when both have the same audio scheme such as stereophonic sound
Solution Approach 1:
The patent merges multiple detection approaches by combining audio scheme analysis with scene change detection. The system uses both the audio-based CM detection unit and the scene change detection unit, integrating their results to identify CM broadcast periods. This combination allows the system to overcome the limitations of audio-only detection and maintain effectiveness even when audio schemes are similar.
Solution Approach 2:
The patent creates a composite detection mechanism that combines different detection methodologies (audio analysis and visual histogram comparison) into a unified CM detection system. This composite approach leverages the strengths of both methods while compensating for their individual weaknesses, particularly when audio schemes alone are insufficient for differentiation.
4Loss of time
If black frame detection is used to identify CM broadcast periods, then CM periods can be detected at boundaries between program main parts and CM broadcasts, but the method is inaccurate when black frames are not inserted or insertion is inconsistent
Solution Approach 1:
The patent changes the detection parameter from relying on the presence/absence of black frames to using histogram difference analysis. Instead of detecting CM broadcasts based on visual black frame insertion, the system analyzes changes in video content characteristics through histogram comparison. This parameter change makes the detection method more robust to variations in black frame insertion practices while maintaining detection speed.
5Measurement precision
If silence period detection is used to identify CM broadcasts, then CM periods can be detected based on intervals between silence periods, but the method cannot accurately identify CM periods when CM lengths are not integer multiples of 15 seconds
Solution Approach 1:
The patent introduces dynamic adjustment of the reference period based on detected silence period intervals. Instead of using a fixed 15-second reference, the system adapts the reference period to match the actual silence interval patterns observed in the video content. This dynamic approach allows accurate CM detection regardless of whether CM lengths are integer multiples of standard time units.
6Ease of operation
If fixed period skipping (e.g., 15 seconds) is used during video viewing, then users can quickly skip through content, but the method cannot accurately skip variable length CM broadcasts
Solution Approach 1:
The patent uses feedback from the detected CM broadcast period information to control the video playback skipping. The system continuously monitors the playback position and compares it with the stored CM period data, adjusting the skip behavior in real-time. This feedback mechanism ensures that skipping operations accurately target CM broadcasts regardless of their variable lengths, while maintaining user-friendly operation.
Data Source
AI summary
There are provided with a video recording apparatus and a scene change extraction method, by which a scene change can be extracted from a video while recording the video. The video recording apparatus 100 includes an encoder 101 for encoding video data PIN, an HDD 105 for storing an encoded stream outputted from the encoder 101, a stream controller 103 for transferring the encoded stream outputted from the encoder 101 to the HDD 105, a scene change extractor 110 for decoding the encoded stream outputted from the encoder 101 and extracting a scene change in the decoded video data, and a recording controller 102 for receiving an encoding-completion notification IEN on a unit-of-encoding basis of the encoded stream from the encoder 101 and outputting a scene change extraction instruction CSC to the scene change extractor 110.