Video Stream Segmentation for Scalable Resolution Adaptation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for transmitting and storing video content with high spatial and temporal resolutions are resource-intensive and inefficient, lacking a solution that balances resource usage with perceived quality and scalability.

Innovation Solution

A method that processes a sequence of images by dividing it into subsequences, applying temporal and spatial subsampling based on image frequency, reducing spatial resolution while maintaining temporal resolution, and inserting the processed subsequences into an output container for efficient encoding and storage, allowing for scalable reconstruction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If variable bitrate compression is used to adapt bit rate to content, then video quality is improved within reduced flow rate range, but the technique is only applicable in a reduced flow rate range

Engineering Contradiction:
Improvevideo qualityVSAvoidflow rate range
Core Design Contradiction:
Manufacturing precisionVSAdaptability or versatility

Solution Approach 1:

The video stream is segmented into multiple layers (base layer and enhancement layers), each with different spatial and temporal resolutions. This segmentation allows the system to adapt to different flow rates by transmitting only the necessary layers, thus expanding the applicable flow rate range while maintaining video quality.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a new dimension of scalability by decomposing video content into multiple spatial and temporal layers. This dimensional decomposition allows flexible adaptation to different network conditions and device capabilities, resolving the contradiction between quality and flow rate range applicability.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Adaptability or versatility

If adaptive bit rate stores several encoded streams with different resolutions, then scalability to different device capabilities is improved, but resource cost increases due to coding, storing and transmitting several versions

Engineering Contradiction:
Improvedevice capability compatibilityVSAvoidresource consumption
Core Design Contradiction:
Adaptability or versatilityVSQuantity of substance

Solution Approach 1:

Instead of storing and transmitting completely separate encoded streams for different resolutions, the patent merges multiple resolutions into a single scalable stream structure. The base layer and enhancement layers are combined in one bitstream, reducing the total quantity of data that needs to be stored and transmitted while maintaining adaptability to different device capabilities.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent implements dynamic scalability where the receiver can dynamically select which layers to decode based on available device resources. This dynamic approach allows a single encoded stream to serve multiple device capabilities without requiring pre-storage of multiple complete versions, thus reducing resource consumption.

Inventive Principle:
Principle #15Dynamics

3Manufacturing precision

If scalable compression produces base layer and enhancement layers, then resolution/quality can be increased with sufficient resources, but implementation complexity increases with large memory size and increased latency

Engineering Contradiction:
Improvespatial and temporal resolutionVSAvoidmemory size and latency
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent applies local quality enhancement by providing different quality levels for different parts of the video stream through layered encoding. The base layer provides essential quality information, while enhancement layers provide additional detail only where needed. This allows receivers to allocate memory and processing resources locally and efficiently, reducing overall system complexity while maintaining high resolution capability when resources are available.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3707900B1Method for forming an output image sequence from an input image sequence, method for reconstructing an input image sequence from an output image sequence, associated devices, server equipment, client equipment and computer programs
Publication Date: 2023.11.29 FOND B COM
  • EP3707900B1 patent drawingFigure 1
  • EP3707900B1 patent drawingFigure 2
  • EP3707900B1 patent drawingFigure 3

AI summary

The invention relates to a method for forming an image sequence, called the output sequence (Ss), from an input image sequence (SE), said input image sequence having an input spatial resolution (RSE) and an input temporal resolution (RTE), said output sequence having an output temporal resolution (RTs) equal to the input temporal resolution (RTE) and an output spatial resolution (RSs) equal to a preset fraction (1/N) of the input spatial resolution (RSE), set by an integer number (N) higher than or equal to 2, characterised in that the method comprises the following steps, which are implemented for a sub-sequence (SSE) of the input sequence, called the current sub-sequence, comprising a preset number (N) of images: obtaining (El) a temporal frequency, called the image frequency (FI), associated with said sub-sequence; processing (E2) the current input sub-sequence to obtain an output sub-sequence (SSs), inserting (E4) the output sub-sequence (SSs) and the associated image frequency (FI) into an output container (Cs).