Immune cell recognition method based on cascade DAE-ResNet

By combining the neural network model based on cascaded DAE-ResNet with traditional Raman spectroscopy with short integral time, the destructive, costly, time-consuming and complex data processing of immune cell detection in the prior art is solved, and fast, lossless and accurate immune cell recognition is achieved, improving analysis efficiency and accuracy.

CN119691514AActive Publication Date: 2025-03-25GUANGDONG UNIV OF TECH
12 Cites 0 Cited by

Patent Information

Application Number
CN202411817413.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-11
Publication Date
2025-03-25
Estimated Expiration
2044-12-11

AI Technical Summary

Technical Problem

The prior art is destructive, expensive, time-consuming and difficult to process complex big data when detecting immune cells, and traditional Raman spectroscopy acquisition requires a long integral time, affecting high-throughput applications.

Method used

A neural network model based on cascaded DAE-ResNet is adopted, combined with traditional Raman spectroscopy with short integral time, to achieve rapid, label-free and accurate recognition of immune cells. This method includes data acquisition, preprocessing, building data sets, building cascading DAE-ResNet network models and training and testing.

Benefits of technology

Fast, non-destructive and accurate immune cell recognition is achieved, the complexity of detection and analysis is reduced, the analysis efficiency and accuracy is improved, and the limitations of the longer integral time required for traditional Raman spectroscopy acquisition are overcome.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

An immune cell recognition method based on cascade DAE-ResNet comprises the following steps: data acquisition: acquiring Raman spectrums of T cells, B cells and NK cells from peripheral blood mononuclear cells by using a Raman spectrometer; the data set is preprocessed, interference signals of the obtained Raman spectrum are removed, the spectrum quality is improved, and a reliable basis is established for follow-up analysis; constructing a data set, dividing the preprocessed Raman spectrums into three categories according to the category of the immune cells, and classifying the Raman spectrums of the same category of cell samples into one category to obtain a spectrum data set; a cascade DAE-ResNet network model is established; and training the cascade DAE-ResNet neural network model. According to the scheme, the single cascade DAE-ResNet neural network is trained to realize the precise recognition of the traditional Raman spectrum immune cells with the integral time of 1 second.
Need to check novelty before this filing date? Find Prior Art