Block-Based PCA Signal Transformation for Lower Computation Cost
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional principal component analysis (PCA) methods face high calculation costs when processing signals with large dimensions, and existing solutions that address this issue often increase hardware costs significantly.
Innovation Solution
A block-based PCA transformation method that divides input signals into one-dimension vectors, calculates a covariance matrix, and uses eigenvectors to obtain projection coefficients, allowing for efficient reconstruction of the original signal with reduced data size by dividing signals into frames or blocks and performing inverse transformations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional PCA processes signals with large dimensions, then complete signal reconstruction is achieved, but calculation cost becomes very high
Solution Approach 1:
The patent divides the input signal into multiple frames or blocks, transforming a large-dimensional signal processing problem into multiple smaller sub-problems. Each frame is processed independently through PCA transformation, reducing the overall calculation burden while maintaining reconstruction accuracy through block-based processing
2Productivity
If hardware solutions are used to reduce calculation cost, then processing efficiency improves, but hardware cost increases significantly
Solution Approach 1:
By segmenting the signal into smaller frames, the patent enables efficient software-based processing that avoids the need for expensive dedicated hardware while maintaining high processing efficiency through reduced computational complexity of each sub-problem
Solution Approach 2:
The patent transforms the signal processing approach by changing the parameter dimension - processing smaller frames sequentially rather than handling the entire large-dimensional signal at once, thereby achieving hardware-efficient processing through parameter optimization
Data Source
AI summary
The present invention provides a block-based principal component analysis transformation method and a device thereof. The principal component analysis transformation method includes: obtaining an input signal; dividing the input signal and obtaining a plurality of one-dimension vectors corresponding to the divided input signal, wherein a number of the one-dimension vectors is a division number; after arranging the one-dimension vectors to a two-dimension vector, subtracting an average value of the one-dimension vectors of the division number to obtain a zero-mean vector; calculating a covariance matrix of the zero-mean vector; calculating an eigenvector of the covariance matrix; multiplying the zero-mean vector by the eigenvector to obtain a projection coefficient.