HDR Image Encoding with Noise Reduction and Layer Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current high dynamic range (HDR) image and video encoding methods, such as JPEG XT, face inefficiencies in data transfer and decoding due to lossy compression, leading to visible compression artefacts and increased data requirements.

Innovation Solution

A method involving lossy encoding, compression-noise reduction using filters like bilateral and anisotropic filters, and encoding difference data to minimize compression noise, thereby reducing artefacts and data size.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If lossy encoding operation is applied to compress image data, then data storage and transfer requirements are reduced, but visible compression artefacts increase and image quality deteriorates

Engineering Contradiction:
Improvedata sizeVSAvoidcompression artefacts
Core Design Contradiction:
Quantity of substanceVSObject-affected harmful factors

Solution Approach 1:

The patent segments the image data into base image layer data and extension image layer data. The base layer contains essential visual information for acceptable quality at lower bitrates, while the extension layer contains additional detail information. This segmentation allows the system to reduce overall data size by selectively encoding different layers with different compression levels, minimizing visible artefacts in the base layer while maintaining option for higher quality with extension layer.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent changes encoding parameters by applying different encoding strengths to different image layers. The base image layer is encoded with parameters optimized for low bitrate performance, while the extension image layer uses parameters that preserve additional detail. This parameter differentiation allows the system to achieve reduced data size without proportionally increasing compression artefacts.

Inventive Principle:
Principle #35Parameter changes

2Manufacturing precision

If JPEG XT encoding with base and extension layers is used, then HDR image quality is improved, but data transfer requirements between encoder and decoder increase

Engineering Contradiction:
Improveimage qualityVSAvoiddata transfer requirements
Core Design Contradiction:
Manufacturing precisionVSQuantity of substance

Solution Approach 1:

The patent implements partial action by providing two tiers of encoding: a base layer that provides sufficient quality for many applications at reduced data transfer, and an extension layer that adds quality for demanding applications. Users can choose to transfer only the base layer data when quality requirements are moderate, or transfer both base and extension layers when high quality is needed, thus optimizing the balance between image quality and data transfer requirements.

Inventive Principle:
Principle #16Partial or excessive action

3Ease of operation

If standard JPEG decoder is used for legacy compatibility, then ease of operation is improved, but HDR reconstruction capability is lost

Engineering Contradiction:
Improvedecoder compatibilityVSAvoidHDR reconstruction
Core Design Contradiction:
Ease of operationVSManufacturing precision

Solution Approach 1:

The patent makes the base image layer data universally compatible with standard JPEG decoders, allowing legacy devices and simple applications to operate the encoded image data without requiring HDR-specific functionality. Meanwhile, HDR-capable decoders can utilize both the base layer and extension layer to reconstruct full HDR images. This universality principle resolves the contradiction by ensuring ease of operation for legacy systems while preserving HDR reconstruction capability for advanced systems.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentUS10491924B2Encoding and decoding of image data
Publication Date: 2019.11.26 ARM LTD
  • US10491924B2 patent drawing
  • US10491924B2 patent drawing
  • US10491924B2 patent drawing

AI summary

A method of encoding image data representing an image is provided. The method includes receiving the image data. The method includes applying a lossy encoding operation to the image data to generate encoded image data representing the image. The method includes decoding the encoded image data to generate decoded image data. The method includes reducing a compression-noise component of the decoded image data to generate reduced-noise decoded image data. The method includes processing the reduced-noise decoded image data and the image data to generate difference data representing a difference between the image data and the reduced-noise decoded image data. The method includes encoding the difference data to generate encoded difference data. A method of decoding image data derived, using a lossy encoding operation, from image data representing an image is also provided. A computing device is also provided.