Combined Prediction Models for Medical Sensor Data Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge in medical data compression is the immense storage and bandwidth requirements due to the vast amounts of data collected from sensors, which existing compression methods struggle to efficiently manage, especially in telemedicine scenarios where real-time access is critical but network bandwidth is insufficient.
Innovation Solution
The technology employs multiple prediction models to accurately predict future data values, combining them into a composite model for efficient compression, using techniques like maximum likelihood estimation and probability mass functions to select the best predictors and compression methods, thereby reducing storage and bandwidth needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple sensors are used to collect medical data at high frequency, then data accuracy and completeness are improved, but storage requirements and network bandwidth needs increase exponentially
Solution Approach 1:
The patent applies preliminary action by performing data compression at the source (patient side) before data transmission and storage. Compression algorithms process the raw sensor data immediately upon collection, reducing the data volume before it enters the transmission and storage pipelines. This preliminary compression prevents the exponential growth of storage requirements while maintaining data accuracy through lossless or controlled lossy compression techniques.
Solution Approach 2:
The patent utilizes parameter changes by dynamically adjusting compression parameters based on data characteristics, sensor types, and clinical priorities. Different compression ratios and algorithms are applied to different sensor data streams, allowing the system to optimize between data fidelity and storage efficiency for each parameter type, thereby managing overall storage requirements effectively.
2Speed
If all collected data is transmitted in real-time, then data availability for diagnosis is improved, but network bandwidth consumption increases beyond available capacity
Solution Approach 1:
The patent implements preliminary compression of data at the patient side before transmission, significantly reducing the data volume that needs to be transmitted over the network. This preliminary action enables real-time or near-real-time data availability for diagnosis while keeping network bandwidth consumption within available capacity limits.
3Measurement precision
If lossless compression is used to maintain complete data fidelity, then data accuracy is improved, but compression efficiency and storage reduction are worsened compared to lossy compression
Solution Approach 1:
The patent applies parameter changes by allowing dynamic selection between lossless and lossy compression modes, and by adjusting compression ratios based on clinical priorities, sensor types, and data characteristics. This flexibility enables the system to optimize compression efficiency for storage and transmission while maintaining adequate data fidelity for diagnostic purposes, rather than universally applying lossless compression.
Data Source
AI summary
Data compression technology (“the technology”) is disclosed that can employ two or more prediction models contemporaneously. The technology receives data from one or more sources; shifts or re-sample one of more corresponding signals; creates a prediction model of uncompressed samples using at least two different individual or composite models; selects a subset of the models for prediction of samples; determines an order in which signals will be compressed; formulates a combined predictions model using the selected subset of models; predicts a future value for the data using the combined compression model; defines a function that has as parameters at least the predicted future values for the data and actual values; selects a compression method for the values of the function; and compresses the data using at least the predicted value of the function.