Video Encoding Device Memory Management for Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video encoding devices face challenges in achieving high-quality encoding with limited memory capacity, particularly in vehicles where memory resources are constrained, leading to degraded image quality due to insufficient reference image storage capacity.
Innovation Solution
A video encoding device that optimizes memory usage by dynamically determining whether to store compressed reference images in a small-capacity memory, using a combination of lossless compression and adaptive intra-inter coding decisions based on image complexity and memory availability, ensuring efficient encoding even with limited memory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If inter-coding is used to achieve higher compression efficiency, then compression ratio is improved, but reference image memory capacity is consumed
Solution Approach 1:
The patent applies dynamics by making the prediction method adaptive rather than fixed. The prediction unit dynamically switches between inter-coding and intra-coding based on real-time evaluation of encoding cost and memory availability. This allows the system to optimize compression efficiency while adapting to varying memory conditions, resolving the contradiction between achieving high compression ratios and preserving reference image memory capacity.
Solution Approach 2:
The patent changes the parameter of prediction mode selection from a static configuration to a dynamic decision based on encoding cost comparison. By evaluating the cost of inter-coding versus intra-coding for each block and selecting the optimal mode, the system adjusts its behavior to maintain compression efficiency while managing memory usage effectively.
2Quantity of substance
If the maximum number of macroblocks for inter-coding is determined in advance based on memory capacity, then memory usage is controlled, but coding efficiency is deteriorated
Solution Approach 1:
The patent replaces the static predetermined macroblock limit with a dynamic selection process. Instead of fixing the number of inter-coded macroblocks in advance, the system evaluates each block individually and determines the prediction mode based on current encoding conditions and memory state. This dynamic approach allows the system to achieve effective memory control while maintaining high coding efficiency.
Solution Approach 2:
The patent performs preliminary evaluation of encoding cost for each block before finalizing the prediction mode decision. By calculating and comparing the costs of inter-coding and intra-coding in advance for each block, the system can make informed decisions that optimize both memory usage and coding efficiency, rather than relying on predetermined limits.
3Quantity of substance
If lossless compression is applied to reference images, then memory storage efficiency is improved, but processing complexity increases
Solution Approach 1:
The patent applies local quality by differentiating the treatment of different blocks based on their specific characteristics. Instead of uniformly applying lossless compression to all reference images, the system evaluates each block's encoding cost and selectively applies prediction modes. This localized approach optimizes memory storage efficiency while avoiding unnecessary processing complexity for blocks where simpler methods suffice.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A video encoding device includes a local decode generation unit for generating a reference image based on a result of encoding of a divided image, a compression unit for compressing the reference image to generate a compressed data, a reference image storage determination unit for determining whether to store the compressed data in a memory, and an inter-prediction unit for performing motion vector search for inter-coding based on a reference image stored in the memory. The reference image storage determination unit sets an allowable data amount used for storing the reference image for each determined area of the moving image data, and determines whether or not to store the compressed data obtained by compressing the reference image in the memory based on the allowable data amount. Inter-prediction unit sets the reference image corresponding to the compressed data stored in the memory as the search range of motion vector search.