Differential Bandwidth Allocation for VR Streaming Regions

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current streaming technologies face challenges in maintaining image quality for virtual reality (VR) content, particularly in VR images that require varying levels of complexity across different regions, as existing methods do not efficiently allocate limited network bandwidth to ensure high-quality reproduction in high complexity regions.

Innovation Solution

The method involves generating segments encoded with various qualities for each region of the content screen based on network conditions and importance information, allowing for adaptive streaming by differentially allocating network bandwidth to prioritize high complexity regions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If encoding is performed with the same quality on all regions regardless of complexity, then network bandwidth is uniformly distributed, but image quality deteriorates in high complexity regions

Engineering Contradiction:
Improveimage qualityVSAvoidnetwork bandwidth
Core Design Contradiction:
Manufacturing precisionVSLoss of energy

Solution Approach 1:

The patent applies local quality by dividing the VR image into multiple regions and encoding each region with different quality levels based on its complexity characteristics. High complexity regions receive higher quality encoding while low complexity regions use lower quality encoding, thereby optimizing overall image quality within limited bandwidth constraints

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent segments the VR image content into multiple regions with different complexity characteristics. By analyzing complexity metrics for each region and separately encoding them, the system achieves differential quality allocation that prevents image quality deterioration in critical high complexity regions

Inventive Principle:
Principle #1Segmentation

2Manufacturing precision

If network bandwidth is uniformly distributed to all regions, then allocation is simple, but image quality deteriorates in high complexity regions requiring high quality reproduction

Engineering Contradiction:
Improveimage quality in high complexity regionsVSAvoidbandwidth allocation complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent performs preliminary analysis of region complexity before encoding by calculating complexity metrics for each region. This pre-analysis enables the system to determine appropriate quality levels for each region in advance, allowing differential quality allocation without complex real-time decision-making during the encoding process

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent changes the encoding quality parameter dynamically based on region complexity characteristics. By adjusting the quality parameter according to pre-calculated complexity metrics, the system achieves adaptive quality allocation that prioritizes high complexity regions while managing overall bandwidth usage

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3528500B1Method and apparatus for allocating differential bandwidth for each screen region by using image complexity information
Publication Date: 2023.06.28 SAMSUNG ELECTRONICS CO LTD
  • EP3528500B1 patent drawingFigure 1
  • EP3528500B1 patent drawingFigure 2
  • EP3528500B1 patent drawingFigure 3

AI summary

Provided are a method and apparatus for allocating differential bandwidth for each screen region by using image complexity information. According to an embodiment, a method and apparatus for receiving streaming data request a server for information relating to content to differentially determine quality of a plurality of regions divided from a predetermined frame of the content according to a network state; among at least one segment for each region encoded with at least one quality for each of the plurality of regions, receive a segment for each region encoded according to the quality for each region determined based on the information relating to the content from the server over a network; generate a reconstruction image for each region by decoding the segment for each region; and merge and render the reconstruction image for each region.