Self-Recursive DCT Circuitry Reducing Multiplicative Complexity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high multiplicative complexity of Discrete Cosine Transform (DCT) calculations leads to inefficient processing and increased resource requirements in digital signal processing, particularly in specialized hardware like digital signal processors (DSPs), which hinders throughput and power efficiency.
Innovation Solution
The implementation of self-recursive DCT techniques using sparse and orthogonal matrices, specifically factorizing DCT II and III transforms into products of simpler matrices, reduces computational complexity by employing even-odd permutations, bidiagonal, and diagonal matrices, enabling fewer multiplications and more efficient hardware implementations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional DCT calculation methods are used, then transform accuracy is maintained, but multiplicative complexity increases leading to inefficient processing
Solution Approach 1:
The DCT calculation is segmented by decomposing the transform into a product of simpler matrices (bidiagonal, diagonal, and orthogonal matrices). This segmentation reduces the multiplicative complexity from O(n²) to O(n log n) by breaking down the complex transformation into manageable stages that can be processed more efficiently
Solution Approach 2:
The invention changes the computational parameters by using self-recursive DCT techniques that reduce the number of multiplications required. By transforming the calculation approach from direct matrix multiplication to a recursive algorithm with fewer multiplicative operations, the processing efficiency improves while maintaining transform accuracy
2Measurement precision
If more multiplication operations are performed, then transform precision is improved, but power consumption increases
Solution Approach 1:
The invention extracts and eliminates redundant multiplication operations from the DCT calculation process. By identifying and removing unnecessary multiplicative steps while retaining the essential transformation operations, the method maintains transform precision with reduced power consumption
Solution Approach 2:
The computational parameters are changed by adopting an algorithm that requires fewer multiplication operations. The self-recursive approach with bidiagonal and diagonal matrix factorization reduces the multiplicative complexity, directly lowering power consumption while preserving the accuracy of the transform results
3Speed
If dedicated DSP hardware with MAC units is used, then processing speed is improved, but device size and resource requirements increase
Solution Approach 1:
The reduced complexity DCT algorithm enables general-purpose processors to perform DCT transformations efficiently without requiring specialized dedicated hardware. The algorithm's reduced multiplicative complexity allows it to run effectively on standard processors, making the capability universal across different device types rather than requiring specialized DSP hardware
Solution Approach 2:
Instead of copying the functionality of dedicated DSP hardware with MAC units, the invention creates a software-based solution that replicates the essential DCT transformation capability through optimized algorithms. This approach achieves similar processing speed improvements without duplicating the complex hardware architecture
Data Source
AI summary
System and techniques for reduced multiplicative complexity discrete cosine transform (DCT) circuitry are described herein. An input data set can be received and, upon the input data set, a self-recursive DCT technique can be performed to produce a transformed data set. Here, the self-recursive DCT technique is based on a product of factors of a specified type of DCT technique. Recursive components of the technique are of the same DCT type as that of the DCT technique. The transformed data set can then be produced to a data consumer.