Method, system, medium and equipment for training model of raw material plasma protein content determination

The fusion of the near-infrared spectral data collected by the dual sensor and the CNN-LSTM model solves the problem of long time to determine raw plasma protein content and temperature interference, achieving rapid and accurate protein content detection.

CN119167997BActive Publication Date: 2025-03-18SHANDONG UNIV
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202411578539.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-11-07
Publication Date
2025-03-18
Estimated Expiration
2044-11-07

AI Technical Summary

Technical Problem

When determining the raw plasma protein content in the prior art, there is a problem that the measurement time is long and the detection needs cannot be met in time. The temperature influence leads to a decrease in the prediction accuracy of the near-infrared spectral analysis model.

Method used

The near-infrared spectrum of raw plasma samples was collected based on different temperatures, and the spectral data containing temperature data were fused with the real protein content through the CNN-LSTM model to eliminate the impact of temperature on the protein content detection results.

Benefits of technology

It realizes rapid and accurate determination of raw plasma protein content, reduces the interference of temperature on spectral analysis, and improves detection efficiency and accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119167997B_ABST
    Figure CN119167997B_ABST
Patent Text Reader

Abstract

The present invention proposes a raw plasma protein content determination model training method, system, medium and equipment, which relates to the field of protein content determination. The method includes: obtaining a raw plasma sample and its true protein content label; using a dual sensor to collect the near-infrared spectrum of the raw plasma sample based on different temperatures to obtain a near-infrared spectrum containing temperature data; training a CNN-LSTM model, inputting the near-infrared spectrum containing temperature data into the CNN-LSTM model, and obtaining a predicted protein content label for each near-infrared spectrum; calculating the loss function of the model according to the predicted protein content label and the true protein content label of each raw plasma sample, and obtaining a trained CNN-LSTM model when the loss of the model is minimized. Based on near-infrared spectroscopy analysis technology and data fusion strategy, the model is used to learn the relationship between spectral data containing temperature information and the true protein content, eliminating the influence of temperature on the protein content detection result.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Spectral analysis method and system based on multistage partial least squares algorithm

    CN113030012A

  • Method for rapid determination of immunoglobulin G content in raw material plasma alcohol precipitation process

    CN104048938A

  • Method for detecting phycocyanin content in cocktail based on LSTM (Long Short Term Memory) and laser-induced fluorescence technology

    CN116660224A