Deep learning-based ICU death risk evaluation system

A deep learning and evaluation system technology, applied in the field of death risk prediction, can solve problems such as inability to assess the risk of death of patients, rough output results, etc., and achieve the effect of close correlation and improved efficiency

Active Publication Date: 2019-01-01
XIAMEN UNIV
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, due to the fixedness of the scoring criteria and the fact that neither of these two scoring criteria utilizes the dynamic characteristics of the patient’s physical signs data, the output results of the existing assessment system are relatively rough and cannot accurately assess the risk of death of the patient. Evaluate

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Deep learning-based ICU death risk evaluation system
  • Deep learning-based ICU death risk evaluation system
  • Deep learning-based ICU death risk evaluation system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0034] Please refer to figure 1 As shown, the present invention discloses a deep learning-based ICU death risk assessment system, which includes an ICU historical database, a first data preprocessing module, a death risk assessment module, a second data preprocessing module, and a human-computer interaction module; The ICU historical database, the first data preprocessing module, the death risk assessment module and the human-computer interaction module are connected in sequence, and the human-computer interaction module, the second data preprocessing module and the death risk assessment module are connected in sequence.

[0035] The historical patient's sign data set and the real final state of the historical patient are stored in the ICU historical database. The first data preprocessing module extracts the historical patient's sign data set from the ICU historical database and performs preprocessing to obtain training sample data, and at the same time extracts the real final...

Embodiment 2

[0058] In this embodiment, the death risk assessment modules are respectively constructed using a one-way LSTM architecture and a two-way LSTM architecture, and the schematic diagrams of the models are as follows image 3 and Figure 4 shown, with figure 2 It can be seen from the comparison that for the models of the two LSTM architectures, the results are only output at the last time step, and the loss value Loss is calculated with the label Label to train the model.

[0059] Such as Figure 5 Shown is the evaluation effect comparison table of the ICU mortality risk assessment system constructed and trained using the unidirectional LSTM architecture (LSTM), the bidirectional LSTM architecture (BiLSTM) and the bidirectional supervised LSTM architecture (BiLSTM-ST). The ICU mortality risk assessment model based on the two-way supervised LSTM architecture is significantly better than that based on LSTM or BiLSTM architecture in terms of Precision (precision rate), Recall (rec...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a deep learning-based ICU death risk evaluation system, which comprises an ICU historical database, a first data preprocessing module, a second data preprocessing module and adeath risk evaluation module, wherein the ICU historical database stores a physical sign data set of a historical patient and the real final state of the historical patient; the first data preprocessing module extracts the physical sign data set of the historical patient in the ICU historical database and carries out preprocessing, training sample data are acquired, and the real final state of thepatient is extracted to give a label for the training sample data; the second data preprocessing module extracts physical sign data of a to-be-evaluated patient inputted by a man-machine interactionmodule and carries out preprocessing; and the death risk evaluation module is built based on a bidirectional supervision-type LSTM neural network. The training sample data and the label value are acquired from the first data preprocessing module for model training, the well-trained model is used to acquire the physical sign data of the to-be-evaluated patient from the second data preprocessing module for evaluation, and finally, the evaluation result is outputted through the man-machine interaction module.

Description

technical field [0001] The invention relates to the field of death risk prediction, in particular to an ICU death risk assessment system based on deep learning. Background technique [0002] At present, the SAPS and APACHE scoring standards are widely used in the hospital ICU to evaluate the risk of death of patients entering the ICU ward. The evaluation system based on these two evaluation methods is based on the data of more than 10 signs in the first 24 hours after the patient enters the ICU. , to assess the patient's physical condition, thereby helping doctors determine more timely and effective treatment measures. However, due to the fixedness of the scoring criteria and the fact that neither of these two scoring criteria utilizes the dynamic characteristics of the patient’s physical signs data, the output results of the existing assessment system are relatively rough and cannot accurately assess the risk of death of the patient. Evaluate. Contents of the invention ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G16H50/20G16H50/30G16H50/70
CPCG16H50/20G16H50/30G16H50/70
Inventor 范晓亮朱耀史佳王程陈龙彪
Owner XIAMEN UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap