Data Compression Using Distribution Models to Cut Quantization Loss
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data transmission methods using uniform quantizers on non-uniformly distributed data result in significant quantization loss, leading to reduced data transmission efficiency.
Innovation Solution
Determine a target data distribution model from a set of models based on the to-be-compressed data, calculate the quantization loss, and use a uniform quantizer to process the data uniformly distributed data to minimize quantization loss.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a uniform quantizer is used to quantize non-uniformly distributed data, then the quantization process is simple and fast, but significant quantization loss occurs
Solution Approach 1:
The patent applies preliminary action by performing data distribution modeling and transformation before the quantization process. Specifically, the system determines the data distribution model, transforms the data to match the model characteristics, and only then applies the uniform quantizer. This preliminary preparation eliminates the quantization loss that would otherwise occur when directly quantizing non-uniform data, while maintaining the simplicity and speed of uniform quantization.
2Loss of information
If the data distribution model is determined through multiple models and calculations, then quantization loss is minimized, but the processing complexity increases
Solution Approach 1:
The patent applies parameter changes by transforming the data parameters to match the determined data distribution model. Instead of changing the quantization method, the system changes the data parameters through transformation operations that align the data characteristics with the model. This approach minimizes quantization loss while avoiding the need for complex adaptive quantization algorithms, thus controlling processing complexity.
Data Source
AI summary
A data transmission method and apparatus are disclosed, to help reduce quantization loss of data that is not uniformly distributed, thereby improving data transmission efficiency. The method includes: selecting a target data distribution model from a plurality of data distribution models based on to-be-compressed data, the plurality of data distribution models, and a first uniform quantizer; determining target compressed data based on the target data distribution model; and sending the target compressed data and first information to a receive end, where the first information includes information identifying the target data distribution model.