Embedded Codec Refinement Bit Allocation for Image Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image and video compression techniques often result in visible artifacts and inefficiencies due to equal allocation of refinement bits across image sub-blocks of varying visual quality, leading to sub-optimal memory usage and compression efficiency in on-chip codecs.
Innovation Solution
An embedded codec circuitry that allocates refinement bits based on visual quality measures, prioritizing areas with lower texture as they are more error-tolerant, allowing for adaptive refinement bit allocation to improve image quality and compression efficiency without increasing memory usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If equal number of refinement bits are allocated in each encoded image sub-block, then the encoding process is simple and deterministic, but visible image artifacts appear and compression efficiency deteriorates
Solution Approach 1:
The patent applies local quality by allocating refinement bits differently across different image sub-blocks based on their visual importance. Regions with low texture complexity receive more refinement bits while high texture regions receive fewer bits, optimizing image quality where it matters most without uniformly increasing complexity across the entire image.
Solution Approach 2:
The patent changes the parameter of refinement bit allocation from a fixed equal distribution to a dynamic distribution based on texture complexity metrics. By calculating texture complexity for each sub-block and adjusting refinement bit allocation accordingly, the system adapts the encoding parameters to local image characteristics, reducing visible artifacts while maintaining compression efficiency.
2Ease of operation
If equal number of refinement bits are allocated in each encoded image sub-block, then the allocation process is straightforward, but compression efficiency becomes sub-optimal
Solution Approach 1:
The patent applies preliminary action by calculating texture complexity metrics for each image sub-block before the refinement bit allocation stage. This pre-calculation allows the system to determine the optimal number of refinement bits needed for each region in advance, enabling efficient compression without requiring complex real-time decisions during the encoding process.
3Device complexity
If equal number of refinement bits are allocated in each encoded image sub-block, then the encoding algorithm is simple to implement, but on-chip memory usage increases
Solution Approach 1:
The patent reduces on-chip memory usage by applying local quality assessment to determine refinement bit allocation. By identifying and prioritizing refinement needs only in regions where they provide the most visual benefit (low texture regions), the system avoids allocating refinement bits uniformly across the entire image, thereby reducing the total memory buffer size required while maintaining encoding algorithm simplicity.
Data Source
AI summary
An embedded codec (EBC) circuitry includes encoder circuitry to encode a plurality of sub-blocks of an image block by an entropy coding scheme to generate a plurality of encoded data blocks. Each encoded data block includes a first plurality of bit-planes and a second plurality of bit-planes. The first plurality of bit-planes include a plurality of entropy coded bits. The encoder circuitry determines a count of refinement bits of a plurality of refinement bits, for an encoded data block of the plurality of encoded data blocks, based on a quality measure of the plurality of encoded data blocks. The quality measure represents a count of the plurality of entropy coded bits in each encoded data block. The encoder circuitry allocates the count of refinement bits in the second plurality of bit-planes of the encoded data block.