Macroblock Scene Change Detection for Video Encoding Efficiency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing scene change detection algorithms in video encoding are inefficient as they fail to recognize subtle changes in video frames, leading to unnecessary coding of frames as independent I-frames, causing inefficiency in compression.

Innovation Solution

A macroblock-based scene change detection algorithm that uses window filtering with weighting factors to determine the degree of difference between current and previous frames, allowing for efficient classification of frames as scene change or non-scene change, thereby optimizing encoding decisions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If existing scene change detection algorithms are used, then scene changes can be detected, but they fail to recognize subtle changes leading to unnecessary I-frame encoding and reduced compression efficiency

Engineering Contradiction:
Improvescene change detection accuracyVSAvoidcompression efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The video frame is divided into multiple macroblocks, and the detection algorithm processes each macroblock independently using window filtering. This segmentation allows for localized analysis of subtle changes without requiring full-frame processing, improving detection accuracy while maintaining computational efficiency for compression.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The algorithm applies weighting factors to different regions within macroblocks, giving different importance to different parts of the frame. This local quality approach enables the system to focus computational resources on areas with subtle changes, improving detection precision without proportionally increasing overall processing load.

Inventive Principle:
Principle #3Local quality

2Productivity

If more accurate scene change detection is implemented, then coding efficiency improves, but the complexity of the detection algorithm increases

Engineering Contradiction:
Improvecoding efficiencyVSAvoiddetection algorithm complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

By segmenting the frame into macroblocks and processing them independently with simple window filtering operations, the algorithm achieves accurate scene change detection through distributed simple computations rather than a single complex global analysis, improving coding efficiency while controlling algorithmic complexity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The algorithm uses adjustable weighting factors and window sizes that can be modified to balance detection accuracy and computational complexity. These parameter changes allow the system to adapt to different video content types and compression requirements without fundamentally changing the algorithm structure.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP2473949B1Scene change detection
Publication Date: 2021.03.10 SONY INTERACTIVE ENTERTAINMENT LLC
  • EP2473949B1 patent drawingFigure 1A~1B
  • EP2473949B1 patent drawingFigure 2A
  • EP2473949B1 patent drawingFigure 2B~5B

AI summary

Scene change detection in encoding digital pictures is disclosed. A statistical quantity µ M is calculated for a given section in a current picture. A window of one or more sections is defined around a co-located section in a previous picture. A statistical sum E is calculated over the sections in the window. A difference between the statistical sum E and the statistical quantity µ M is calculated. The difference between E and µ M is used to determine whether the given section is a scene-change section. Whether the current picture is a scene-change picture may be determined from the number of scene change sections. Information indicating whether or not the current picture is a scene-change picture may be stored or transferred.