Scene-Adaptive Video Encoding for Frame Loss Resilience
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding technologies lack flexibility in reference frame structures, leading to inflexible encoding processes and reduced efficiency, particularly in handling frame losses due to fixed reference frame structures like IPPP, which affect decoding and reconstruction.
Innovation Solution
Adaptive switching of reference frame structures based on scene detection and network conditions to determine optimal reference frames and encoding layers, allowing for flexible encoding and improved efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a fixed reference frame structure (e.g., IPPP) is used for video encoding, then the encoding process is simple and consistent, but the flexibility is low and frame loss resilience is poor
Solution Approach 1:
The patent implements dynamic switching between different reference frame structures (IPPP, PBPB, hierarchical) based on scene detection results. The encoder transitions from a static, fixed reference frame structure to a dynamic one that adapts to scene changes, thereby improving flexibility and frame loss resilience without permanently increasing system complexity
Solution Approach 2:
The patent changes the reference frame structure parameters (which frames serve as reference frames and their temporal distances) based on detected scene status. When scene changes are detected, the system adjusts parameters such as reference frame selection and temporal layering to optimize encoding performance for different scene conditions
2Productivity
If scene detection and adaptive reference frame switching are implemented, then encoding efficiency and frame loss resilience are improved, but the encoding process complexity increases
Solution Approach 1:
The patent performs scene detection in advance before encoding each frame to determine the appropriate reference frame structure. This preliminary analysis allows the encoder to proactively select optimal encoding parameters, improving efficiency by avoiding suboptimal fixed structures while managing complexity through structured detection algorithms
Solution Approach 2:
The system uses scene detection results as feedback to continuously adjust the reference frame structure during encoding. This closed-loop approach optimizes encoding efficiency by adapting to actual scene conditions while managing complexity through systematic feedback processing and structured decision-making
3Reliability
If every frame is used as a reference frame in IPPP structure, then encoding consistency is maintained, but decoding and reconstruction are affected when frame loss occurs
Solution Approach 1:
The patent prepares multiple reference frame structures in advance (IPPP, PBPB, hierarchical) with different temporal layering and reference frame selections. When frame loss is detected or anticipated, the system can switch to a pre-prepared alternative structure that is more resilient to the specific loss pattern, cushioning the impact on decoding reliability
Solution Approach 2:
The system dynamically changes reference frame parameters (which frames are used as references and their temporal distances) based on scene status and loss conditions. By adjusting these parameters, the encoder creates more robust reference structures that maintain decoding reliability under frame loss conditions
Data Source
AI summary
A method includes performing scene detection on a current frame of picture to obtain a scene status of the current frame of picture; determining, based on the scene status, a reference frame structure corresponding to the current frame of picture, where the reference frame structure indicates a reference frame of picture of the current frame of picture and an encoding layer of the current frame of picture; and encoding the current frame of picture into a bit stream based on the reference frame structure. In a process of encoding a video, a reference frame of picture and an encoding layer of each frame of picture are adjusted in real time with reference to features such as whether scene switching occurs or whether a scene is kept stable in each frame of picture.