Rectangular Block Dequantization Algorithm for Video Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video compression technologies are optimized for square blocks of coefficients, failing to effectively address the need for quantization and dequantization of rectangular blocks, which are essential for efficient video media compression in applications like videoconferencing and internet streaming.
Innovation Solution
The implementation of specific quantization and dequantization algorithms for rectangular blocks of coefficients, using different algorithms based on the square root of the block size, with separate sets of constants and mapping tables for integer and non-integer values, to generate and reconstruct pictures efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If quantization and dequantization algorithms are optimized for square blocks, then processing efficiency for square blocks is improved, but adaptability to rectangular blocks deteriorates
Solution Approach 1:
The patent changes the algorithm parameters and constants based on the block shape (square vs. rectangular). Different sets of constants are selected depending on whether the block dimensions satisfy the square block condition, allowing the system to maintain high processing efficiency for square blocks while extending support to rectangular blocks through parameter adaptation.
Solution Approach 2:
The quantization and dequantization process becomes dynamic by selecting different algorithms based on the block size characteristics. The system adapts its behavior at runtime based on whether the input block is square or rectangular, enabling versatile processing without requiring separate static algorithms for each block type.
2Measurement precision
If separate algorithms are used for square and rectangular blocks, then processing accuracy is improved, but device complexity increases
Solution Approach 1:
The patent segments the quantization and dequantization process into distinct paths based on block shape. By dividing the processing logic into separate handling for square blocks and rectangular blocks, the system achieves high processing accuracy for each block type while organizing the complexity in a manageable, structured manner through clear decision points and separate constant sets.
Data Source
AI summary
A system includes an encoder that performs quantization on a rectangular block of quantized transform coefficients and a decoder that performs dequantization on a rectangular block of transform coefficients. The decoder is configured to receive a W×H block of quantized transform coefficients, wherein W is a row width of the block and H is a column height of the block. The decoder is further configured to apply dequantization processing to the W×H block of quantized transform coefficients to generate a block of transform coefficients, which are used to generate a picture corresponding to the W×H block of quantized transform coefficients. The dequantization processing comprises applying a first dequantization algorithm when an output M of a function M=log2(√{square root over (W·H)}) is equal to an integer value, and applying a second dequantization algorithm when the output M is equal to a non-integer value.