Video Encoding Reference Image Selection for Screen Content
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression technologies, such as HEVC, face challenges in optimizing the use of reference images for efficient encoding and decoding, particularly in combining intra-layer and inter-layer prediction with scalability and screen content features, which affects compression efficiency and quality.
Innovation Solution
A method for encoding and decoding video streams that selects between reconstructed spatial and low-resolution portions of images as reference images based on a control parameter, allowing for the use of interlayer reference images to enhance prediction accuracy and efficiency, while managing memory usage and bit-rate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If interlayer reference images are used for prediction, then prediction accuracy is improved, but memory usage increases
Solution Approach 1:
The patent segments the reference image management into two distinct parts: intra-layer reference images (current layer) and inter-layer reference images (lower layers). This segmentation allows selective management of reference images based on layer type, enabling the system to optimize memory usage by only storing and accessing inter-layer references when needed for enhancement layers, while base layers use only intra-layer references.
Solution Approach 2:
The patent implements dynamic control over the use of inter-layer reference images through syntax elements that can be conditionally activated. The system dynamically adjusts reference image selection based on picture type (intra/inter), layer type (base/enhancement), and configuration flags, allowing flexible memory management that adapts to different encoding scenarios and quality requirements.
2Productivity
If multiple reference images are maintained for both intra-layer and inter-layer prediction, then compression efficiency is improved, but device complexity increases
Solution Approach 1:
The patent applies different reference image management strategies to different parts of the video coding system: base layers use simplified intra-layer reference management, while enhancement layers utilize inter-layer reference management. This local differentiation allows each layer type to use the most appropriate reference strategy, optimizing compression efficiency without uniformly increasing complexity across the entire system.
Solution Approach 2:
The patent creates a universal reference image management framework that handles both intra-layer and inter-layer prediction through a unified syntax structure. The same reference picture list mechanisms and prediction algorithms are used across all layers, with flexible activation based on configuration flags, reducing overall system complexity by avoiding separate dedicated mechanisms for each prediction type.
3Manufacturing precision
If IBC mode is prioritized over interlayer reference images, then screen content coding quality is improved, but adaptability to different content types decreases
Solution Approach 1:
The patent implements dynamic switching between IBC mode and inter-layer prediction mode based on content type detection and configuration flags. The system can adaptively select the most appropriate prediction method for each picture or coding unit, allowing prioritization of IBC for screen content while maintaining the option to use inter-layer references for natural video content, thus balancing quality and adaptability.
Solution Approach 2:
The patent uses profile-specific configuration parameters and syntax flags to control the behavior of reference image selection. Different profiles (e.g., screen content vs. natural video) can set different default values for these parameters, allowing the same underlying mechanism to optimize for different content types without reducing adaptability. The parameters control aspects such as mandatory IBC usage, inter-layer reference availability, and prediction mode selection.
Data Source
AI summary
The present invention relates to encoding an image of a video stream according to at least one coding mode selected among a plurality of coding modes used to encode images of the video stream, where blocks of the image to be encoded are predicted as a function of at least one reference image from a set of at least one reference image, the at least one reference image comprising at least a reconstructed spatial portion of the image to be encoded and at least a low resolution portion of the image to be encoded, the choice between a reconstructed spatial sub-portion or a low resolution sub-portion to be used for encoding a block of the image to be encoded being determined as a function of a control parameter.