Disclosure of Invention
In view of the above problems, the present invention aims to solve the problems that the existing hyperbaric oxygen chamber and the existing sub-low temperature mattress are both used as an independent treatment means for ischemic and hypoxic brain injury, the curative effect is single, the setting parameters of the hyperbaric oxygen chamber and the sub-low temperature mattress are mainly set by the experience of medical doctors, the physiological parameters of the neonate such as electroencephalogram, blood oxygen concentration, heart rate and the like are not monitored in the treatment process, and the safety in the treatment process is low.
The embodiment of the invention provides a neonate intelligent regulation and control system based on combination of a sub-low temperature mattress and a hyperbaric oxygen chamber, which comprises: the system comprises a sub-low temperature bed, a hyperbaric oxygen chamber, a physiological information monitoring module and an intelligent regulation and control module; wherein the sub-cryobed is arranged inside the hyperbaric oxygen chamber;
the physiological information monitoring module is arranged in the hyperbaric oxygen chamber and is used for collecting multiple physiological information parameters of the neonate and transmitting the multiple physiological information parameters of the neonate to the intelligent control module;
the intelligent control module is connected with the sub-low temperature bed, the hyperbaric oxygen chamber and the physiological information monitoring module, and is used for generating a newborn brain functional state by using an intelligent control algorithm based on a plurality of physiological information parameters of the newborn and controlling the setting parameters of the sub-low temperature bed and the hyperbaric oxygen chamber based on the newborn brain functional state.
In one embodiment, the physiological information monitoring module includes: the electroencephalogram monitoring unit, the electrocardiogram monitoring unit and the near infrared spectrum monitoring unit;
the electroencephalogram monitoring unit is used for acquiring electroencephalogram signals and generating amplitude integrated electroencephalograms based on the electroencephalogram signals;
the electrocardiogram monitoring unit is used for collecting electrocardiogram data of the neonate and extracting electrocardiogram data characteristics;
the near infrared spectrum monitoring unit is used for collecting near infrared spectrum signals and generating the blood oxygen content of the brain of the newborn based on the near infrared spectrum signals.
In one embodiment, the intelligent regulation module comprises: the device comprises an amplitude integration electroencephalogram feature extraction unit, a brain function feature generation unit and a newborn brain function state generation unit;
the amplitude integrated electroencephalogram feature extraction unit is used for performing convolution pooling on the amplitude electroencephalogram by using a convolution neural network and extracting convolution integrated electroencephalogram features;
the brain function feature generation unit is connected with the amplitude integration electroencephalogram feature extraction unit and is used for fusing the convolution integration electroencephalogram feature, the electrocardiographic data feature and the blood oxygen content of the neonatal brain to generate a brain function feature;
the newborn brain function state generating unit is connected with the brain function feature generating unit and used for inputting the brain function features into a full connection layer for classification, recognizing the newborn brain function state by utilizing a convolutional neural network and regulating and controlling the setting parameters of the sub-low temperature bed and the hyperbaric oxygen chamber based on the newborn brain function state.
In one embodiment, the neonatal brain functional state comprises:
dangerous state, severe ischemia-hypoxia brain function damage, mild ischemia-hypoxia brain damage and normal brain function state.
In one embodiment, the setting parameters of the sub-cryobed and the hyperbaric oxygen chamber comprise:
pressure, oxygen concentration, and sub-cryogenic set temperature.
In view of the above, in a second aspect of the present application, there is also provided a method for intelligently regulating a newborn baby based on the combination of a sub-low temperature mattress and a hyperbaric oxygen chamber, comprising:
the physiological information monitoring module collects multiple physiological information parameters of the neonate and transmits the multiple physiological information parameters of the neonate to the intelligent control module;
the intelligent control module generates a newborn brain function state by using an intelligent control algorithm based on the multiple physiological information parameters of the newborn, and controls the setting parameters of the sub-low temperature bed and the hyperbaric oxygen chamber based on the newborn brain function state.
In one embodiment, the physiological information monitoring module collects multiple physiological information parameters of a neonate and transmits the multiple physiological information parameters of the neonate to the intelligent control module, and the method comprises the following steps:
the electroencephalogram monitoring unit acquires an electroencephalogram signal and generates an amplitude integrated electroencephalogram based on the electroencephalogram signal;
the method comprises the following steps that an electrocardio monitoring unit collects electrocardio data of a newborn and extracts electrocardio data characteristics;
the near infrared spectrum monitoring unit collects near infrared spectrum signals and generates the blood oxygen content of the brain of the newborn based on the near infrared spectrum signals.
In one embodiment, the intelligent control module generates a brain function state of the neonate by using an intelligent control algorithm based on a plurality of physiological information parameters of the neonate, and controls setting parameters of the sub-hypothermia bed and the hyperbaric oxygen chamber based on the brain function state of the neonate, including:
the amplitude integration electroencephalogram feature extraction unit performs convolution pooling on the amplitude electroencephalogram by using a convolution neural network, and extracts convolution integration electroencephalogram features;
the brain function feature generation unit fuses the convolution integration electroencephalogram feature, the electrocardio data feature and the blood oxygen content of the newborn brain to generate a brain function feature;
the brain function state generation unit inputs the brain function features into the full-connection layer for classification, generates a brain function state of the neonate by using a convolutional neural network, and regulates and controls the setting parameters of the sub-low temperature bed and the hyperbaric oxygen chamber based on the brain function state of the neonate.
In one embodiment, the neonatal brain functional state comprises:
dangerous state, severe ischemia-hypoxia brain function damage, mild ischemia-hypoxia brain damage and normal brain function state.
In one embodiment, the setting parameters of the sub-cryobed and the hyperbaric oxygen chamber comprise:
pressure, oxygen concentration, and sub-cryogenic set temperature.
The technical scheme provided by the embodiment of the invention has the beneficial effects that at least:
according to the intelligent regulation and control system for the neonate based on the combination of the sub-low temperature mattress and the hyperbaric oxygen chamber, the hyperbaric oxygen chamber is combined with the sub-low temperature mattress, and the flexible switching of three treatment modes of single hyperbaric oxygen treatment, single sub-low temperature treatment and hyperbaric oxygen and sub-low temperature combined treatment is realized. And, when the neonate carries on hyperbaric oxygen treatment or inferior hypothermia treatment, carry on the brain function related many physiological parameter monitoring to the patient, when the neonate can't self-expression change and feeling in the course of treatment, carry on the brain function monitoring in the course of treatment to the neonate, can detect whether the treatment has curative effect objectively on the basis of the monitoring data. Finally, the intelligent regulation and control algorithm for hyperbaric oxygen therapy and sub-hypothermia therapy combined with physiological information monitoring utilizes a convolutional neural network, the characteristics are automatically extracted for partial physiological information, various physiological parameters are subjected to characteristic fusion, finally, the fused characteristics are utilized for classification and identification, the brain function state of the neonate is identified, the parameter settings of a hyperbaric oxygen chamber and a sub-hypothermia mattress are intelligently regulated and controlled according to different brain injury degrees and recommended parameters of different injury degrees which are preset by experts in the field of brain injury, the whole body oxygen supply can be efficiently improved for the neonate, meanwhile, the condition that the neonate is seriously uncomfortable and not known in the treatment process can be avoided, the alarm can be timely carried out on dangerous conditions, and the safety in the treatment process is enhanced.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
Referring to fig. 1, an embodiment of the present invention provides a neonate intelligent regulation and control system based on a combination of a sub-low temperature mattress and a hyperbaric oxygen chamber, including: the system comprises a sub-low temperature bed 1, a hyperbaric oxygen chamber 2, a physiological information monitoring module 3 and an intelligent regulation and control module 4; wherein the sub-low temperature bed 1 is arranged inside the hyperbaric oxygen chamber 2;
the physiological information monitoring module 3 is arranged in the hyperbaric oxygen chamber 2 and is used for collecting a plurality of physiological information parameters of the neonate and transmitting the plurality of physiological information parameters of the neonate to the intelligent control module 4.
Specifically, the newborn with ischemic and hypoxic brain injury is subjected to multiple physiological information parameter monitoring during hyperbaric oxygen or sub-hypothermia treatment so as to evaluate brain functions (mainly including but not limited to amplitude integrated electroencephalogram (aEEG) and original electroencephalogram (EEG) monitoring, Near Infrared Spectroscopy (NIRs) monitoring and Electrocardiogram (ECG) monitoring).
The intelligent control module 4 is connected with the sub-low temperature bed 1, the hyperbaric oxygen chamber 2 and the physiological information monitoring module 3, and is used for generating a newborn brain functional state by using an intelligent control algorithm based on a plurality of physiological information parameters of the newborn, and controlling the setting parameters of the sub-low temperature bed 1 and the hyperbaric oxygen chamber 2 based on the newborn brain functional state.
Specifically, the setting parameters of the sub-low temperature bed 1 and the hyperbaric oxygen chamber 2 include: pressure, oxygen concentration, and sub-cryogenic set temperature.
Furthermore, setting parameter schemes of the sub-hypothermia bed 1 and the hyperbaric oxygen chamber 2 are set as b, c and d by experts in the field of neonatal ischemic-hypoxic brain injury aiming at different degrees of ischemic-hypoxic brain injury in early stage; the parameter setting scheme of the sub-low temperature bed 1 and the hyperbaric oxygen chamber 2 is a parameter setting scheme that b represents the serious brain injury, the parameter setting scheme of the sub-low temperature bed 1 and the hyperbaric oxygen chamber 2 is a parameter setting scheme that c represents the light brain injury, and the parameter setting scheme of the sub-low temperature bed 1 and the hyperbaric oxygen chamber 2 is a parameter setting scheme that d represents the normal brain function state.
In the embodiment, the hyperbaric oxygen chamber is combined with the sub-low temperature bed, so that three treatment modes of separate hyperbaric oxygen treatment, separate sub-low temperature treatment and hyperbaric oxygen and sub-low temperature combined treatment are flexibly switched. And, when the neonate carries on hyperbaric oxygen treatment or inferior hypothermia treatment, carry on the brain function related many physiological parameter monitoring to the patient, when the neonate can't self-expression change and feeling in the course of treatment, carry on the brain function monitoring in the course of treatment to the neonate, can detect whether the treatment has curative effect objectively on the basis of the monitoring data. Finally, the intelligent regulation and control algorithm for hyperbaric oxygen therapy and sub-hypothermia therapy combined with physiological information monitoring utilizes a convolutional neural network, the characteristics are automatically extracted for partial physiological information, various physiological parameters are subjected to characteristic fusion, finally, the fused characteristics are utilized for classification and identification, the brain function state of the neonate is identified, the parameter settings of a hyperbaric oxygen chamber and a sub-hypothermia mattress are intelligently regulated and controlled according to different brain injury degrees and recommended parameters of different injury degrees which are preset by experts in the field of brain injury, the whole body oxygen supply can be efficiently improved for the neonate, meanwhile, the condition that the neonate is seriously uncomfortable and not known in the treatment process can be avoided, the alarm can be timely carried out on dangerous conditions, and the safety in the treatment process is enhanced.
It should be noted that, referring to fig. 2, the hyperbaric oxygen chamber 2 is combined with the sub-cryobed 1, and the improved hyperbaric oxygen chamber can realize three modes of treatment for the patient: separate hyperbaric oxygen chamber treatment, separate sub-hypothermia treatment, hyperbaric oxygen chamber and sub-hypothermia synchronous treatment; the improved hyperbaric oxygen chamber can be flexibly switched among the three treatment modes, and when the existing single hyperbaric oxygen chamber and the single sub-hypothermia bed provide treatment services, a newborn needs to be moved, the treatment mode is not flexible to switch, and synchronous treatment of the hyperbaric oxygen chamber and the sub-hypothermia cannot be realized.
In one embodiment, the physiological information monitoring module 3 includes: an electroencephalogram monitoring unit 5, an electrocardiogram monitoring unit 6 and a near infrared spectrum monitoring unit 7;
the electroencephalogram monitoring unit 5 is used for collecting electroencephalogram signals and generating amplitude integrated electroencephalograms based on the electroencephalogram signals.
Specifically, the electroencephalogram signal is collected, the collected electroencephalogram signal is converted into an amplitude-integrated electroencephalogram (aag) and an original electroencephalogram (EEG), and the amplitude-integrated electroencephalogram and original electroencephalogram monitoring technology is applied to the improved hyperbaric oxygen chamber 2 (i.e. the hyperbaric oxygen chamber 2 provided with the sub-hypothermic bed 1), so that the brain function state of the neonate can be monitored in real time.
Furthermore, the amplitude integrated electroencephalogram (aEEG) is a simplified electroencephalogram monitoring technology, the electroencephalogram information can be collected by adopting fewer electrodes, the anti-interference capability is strong, the brain function damage degree can be easily judged by clinicians, and the amplitude integrated electroencephalogram (aEEG) is suitable for early bedside brain function monitoring and is widely applied to the fields of newborn brain development maturity evaluation, newborn brain damage detection, newborn seizure monitoring and the like together with the original electroencephalogram (EEG).
The electrocardiogram monitoring unit 6 is used for collecting electrocardiogram data of the neonate and extracting characteristics of the electrocardiogram data.
Specifically, electrocardiographic monitoring is used in the improved hyperbaric oxygen chamber 2, and after acquiring electrocardiographic data of a newborn, electrocardiographic data features (mainly including but not limited to Heart Rate (HR), P wave time, R peak time, PR interval, QRS interval, RV5/SV1, and ST segment average amplitude) are extracted.
Furthermore, the electrocardiographic monitoring can monitor vital signs of the neonate, and can monitor the heart rate, heart rhythm, myocardial blood supply condition and electrolyte disorder of the neonate.
The near infrared spectrum monitoring unit 7 is used for collecting near infrared spectrum signals and generating the blood oxygen content of the newborn brain based on the near infrared spectrum signals.
Specifically, Near Infrared Spectra (NIRs) can monitor the hemoglobin concentration (HbO2) and the blood oxygen saturation (SaO2) of the brain tissue of the newborn in a noninvasive and real-time manner, and the near infrared spectrum monitoring is applied to the improved hyperbaric oxygen chamber 2, so that the blood oxygen content of the brain of the newborn is monitored, the degree of brain injury progress is judged, and the treatment effects of the hyperbaric oxygen chamber 2 and the sub-cryo-bed 1 can be fed back for treatment.
In one embodiment, referring to fig. 3, the intelligent control module 4 includes: an amplitude integration electroencephalogram feature extraction unit 8, a brain function feature generation unit 9 and a newborn brain function state generation unit 10;
the amplitude-integrated electroencephalogram feature extraction unit 8 is used for performing convolution pooling on the amplitude electroencephalogram by using a convolution neural network to extract convolution-integrated electroencephalogram features.
Specifically, the convolution is performed on the input amplitude electroencephalogram, then the Relu function is used for activation, the convolution integrated electroencephalogram feature map is extracted, and the calculation formula is as follows:
in the above formula, the first and second carbon atoms are,
representing a convolution integration electroencephalogram characteristic output diagram of a previous layer;
representing a convolution kernel; m
jRepresenting the jth input amplitude electroencephalogram;
representing the bias of the convolution integrated electroencephalogram feature map; f (.) denotes the Relu activation function;
and (4) representing the j-th convolution integrated electroencephalogram characteristic diagram of the l layers.
Further, pooling (down-sampling) is performed on the convolution integrated electroencephalogram feature map, and a specific calculation formula is as follows:
in the above formula, the first and second carbon atoms are,
weight coefficients representing down-sampling; down (.) denotes a downsampling function;
and (4) representing a convolution integration electroencephalogram characteristic output diagram of the previous layer.
The brain function feature generation unit 9 is connected to the amplitude-integrated electroencephalogram feature extraction unit 8, and is configured to fuse the convolution-integrated electroencephalogram features, the electrocardiographic data features, and the blood oxygen content of the neonatal brain to generate brain function features.
Specifically, a two-dimensional convolution integration electroencephalogram feature map is converted into a one-dimensional feature map, then features of electroencephalogram data and features of cerebral blood oxygen data are fused, and a calculation formula is as follows:
in the above formula, the first and second carbon atoms are,
the characteristics of the brain function are shown,
the convolution is expressed to integrate the characteristics of the brain electricity,
representing the blood oxygen content of the newborn brain,
representing the electrocardiographic data.
The newborn brain function state generating unit 10 is connected with the brain function feature generating unit 9 and used for inputting the brain function features into a full connection layer for classification, recognizing the newborn brain function state by using a convolutional neural network, and regulating and controlling the setting parameters of the sub-low temperature bed 1 and the hyperbaric oxygen chamber 2 based on the newborn brain function state.
Specifically, the brain function features are input into the full-link layer for classification, and the calculation formula is as follows:
in the above formula, ωlA weight coefficient representing a full connection layer; blRepresenting the bias of the fully connected layer.
Specifically, the brain function state of the newborn comprises: dangerous state, severe ischemia-hypoxia brain function damage, mild ischemia-hypoxia brain damage and normal brain function state.
Further, referring to fig. 4, the brain function status of the newborn is identified by a convolutional neural network, and the identification results are a, b, c and d; if the identification result is a, the newborn brain function state generating unit 10 automatically and immediately calls medical care personnel corresponding to the identification result as a dangerous state; if the identification result is b, the corresponding identification result is that the ischemic and hypoxic brain injury is serious, and the newborn brain function state generating unit 10 automatically changes the setting parameters (such as pressure, oxygen concentration, sub-hypothermia setting temperature and the like) of the sub-hypothermia bed 1 and the hyperbaric oxygen chamber 2 into a parameter setting scheme b when the brain injury is serious; if the identification result is c, the corresponding identification result is that the ischemic and hypoxic brain injury is light, and the newborn brain function state generating unit 10 automatically changes the setting parameters (such as pressure, oxygen concentration, sub-low temperature setting temperature and the like) of the sub-low temperature bed 1 and the hyperbaric oxygen chamber 2 into a parameter setting scheme c when the brain injury is light; if the recognition result is d, the newborn brain function state generating unit 10 automatically changes the setting parameters (such as pressure, oxygen concentration, and sub-hypothermia setting temperature) of the sub-hypothermia bed 1 and the hyperbaric oxygen chamber 2 to the parameter setting scheme d when the brain function state is normal, corresponding to the recognition result that the brain function state is normal.
Referring to fig. 5, the intelligent neonatal regulation method based on the combination of the sub-low temperature mattress and the hyperbaric oxygen chamber comprises the following steps:
s501, the physiological information monitoring module collects multiple physiological information parameters of the neonate and transmits the multiple physiological information parameters of the neonate to the intelligent control module.
Specifically, the newborn with ischemic and hypoxic brain injury is subjected to multiple physiological information parameter monitoring during hyperbaric oxygen or sub-hypothermia treatment so as to evaluate brain functions (mainly including but not limited to amplitude integrated electroencephalogram (aEEG) and original electroencephalogram (EEG) monitoring, Near Infrared Spectroscopy (NIRs) monitoring and Electrocardiogram (ECG) monitoring).
S502, the intelligent regulation and control module generates a brain function state of the neonate by using an intelligent regulation and control algorithm based on a plurality of physiological information parameters of the neonate, and regulates and controls setting parameters of the sub-low temperature bed and the hyperbaric oxygen chamber based on the brain function state of the neonate.
Specifically, the setting parameters of the sub-low temperature bed and the hyperbaric oxygen chamber comprise: pressure, oxygen concentration, and sub-cryogenic set temperature.
Furthermore, setting parameter schemes of the sub-hypothermia bed 1 and the hyperbaric oxygen chamber 2 are set as b, c and d by experts in the field of neonatal ischemic-hypoxic brain injury aiming at different degrees of ischemic-hypoxic brain injury in early stage; the parameter setting scheme of the sub-low temperature bed 1 and the hyperbaric oxygen chamber 2 is a parameter setting scheme that b represents the serious brain injury, the parameter setting scheme of the sub-low temperature bed 1 and the hyperbaric oxygen chamber 2 is a parameter setting scheme that c represents the light brain injury, and the parameter setting scheme of the sub-low temperature bed 1 and the hyperbaric oxygen chamber 2 is a parameter setting scheme that d represents the normal brain function state.
It should be noted that the hyperbaric oxygen chamber is combined with the sub-cryobed, and the improved hyperbaric oxygen chamber can realize three modes of treatment for patients: separate hyperbaric oxygen chamber treatment, separate sub-hypothermia treatment, hyperbaric oxygen chamber and sub-hypothermia synchronous treatment; the improved hyperbaric oxygen chamber can be flexibly switched among the three treatment modes, and when the existing single hyperbaric oxygen chamber and the single sub-hypothermia bed provide treatment services, a newborn needs to be moved, the treatment mode is not flexible to switch, and synchronous treatment of the hyperbaric oxygen chamber and the sub-hypothermia cannot be realized.
In one embodiment, as shown in fig. 6, in step S501, the acquiring, by the physiological information monitoring module, a plurality of physiological information parameters of a neonate, and transmitting the plurality of physiological information parameters of the neonate to the intelligent control module includes:
s5011, collecting electroencephalogram signals by an electroencephalogram monitoring unit, and generating amplitude integrated electroencephalograms based on the electroencephalogram signals.
Specifically, the electroencephalogram signal is collected and converted into an amplitude-integrated electroencephalogram (aEEG) and an original electroencephalogram (EEG), and the amplitude-integrated electroencephalogram and original electroencephalogram monitoring technology is applied to an improved hyperbaric oxygen chamber (namely, a hyperbaric oxygen chamber provided with a sub-low temperature mattress), so that the brain function state of the newborn can be monitored in real time.
Furthermore, the amplitude integrated electroencephalogram (aEEG) is a simplified electroencephalogram monitoring technology, the electroencephalogram information can be collected by adopting fewer electrodes, the anti-interference capability is strong, the brain function damage degree can be easily judged by clinicians, and the amplitude integrated electroencephalogram (aEEG) is suitable for early bedside brain function monitoring and is widely applied to the fields of newborn brain development maturity evaluation, newborn brain damage detection, newborn seizure monitoring and the like together with the original electroencephalogram (EEG).
S5012, the electrocardiogram monitoring unit collects electrocardiogram data of the neonate and extracts characteristics of the electrocardiogram data.
Specifically, the electrocardiographic monitoring is used in an improved hyperbaric oxygen chamber, and after acquiring electrocardiographic data of a newborn, electrocardiographic data features (mainly including but not limited to Heart Rate (HR), P wave time, R peak time, PR interval, QRS interval, RV5/SV1 and ST segment average amplitude) are extracted.
Furthermore, the electrocardiographic monitoring can monitor vital signs of the neonate, and can monitor the heart rate, heart rhythm, myocardial blood supply condition and electrolyte disorder of the neonate.
S5013, collecting a near infrared spectrum signal by a near infrared spectrum monitoring unit, and generating the blood oxygen content of the brain of the newborn based on the near infrared spectrum signal.
Specifically, Near Infrared Spectra (NIRs) can monitor the hemoglobin concentration (HbO2) and the blood oxygen saturation (SaO2) of the brain tissue of the newborn in a non-invasive and real-time manner, and the near infrared spectrum monitoring is applied to an improved hyperbaric oxygen chamber, so that the blood oxygen content of the brain of the newborn is monitored, the degree of brain injury progress is judged, and the treatment effects of treating the hyperbaric oxygen chamber and a sub-low temperature mattress can be fed back.
In one embodiment, as shown in fig. 7, in step S502, the generating, by the intelligent control module, a brain function state of a neonate based on a plurality of physiological information parameters of the neonate by using an intelligent control algorithm, and controlling setting parameters of the sub-hypothermia bed and the hyperbaric oxygen chamber based on the brain function state of the neonate includes:
s5021, the amplitude integration electroencephalogram feature extraction unit conducts convolution pooling on the amplitude electroencephalogram by means of the convolution neural network, and convolution integration electroencephalogram features are extracted.
Specifically, the convolution is performed on the input amplitude electroencephalogram, then the Relu function is used for activation, the convolution integrated electroencephalogram feature map is extracted, and the calculation formula is as follows:
in the above formula, the first and second carbon atoms are,
representing a convolution integration electroencephalogram characteristic output diagram of a previous layer;
representing a convolution kernel; m
jRepresenting the jth input amplitude electroencephalogram;
representing the bias of the convolution integrated electroencephalogram feature map; f (.) denotes the Relu activation function;
and (4) representing the j-th convolution integrated electroencephalogram characteristic diagram of the l layers.
Further, pooling (down-sampling) is performed on the convolution integrated electroencephalogram feature map, and a specific calculation formula is as follows:
in the above formula, the first and second carbon atoms are,
weight coefficients representing down-sampling; down (.) denotes a downsampling function;
and (4) representing a convolution integration electroencephalogram characteristic output diagram of the previous layer.
S5022, the brain function feature generating unit fuses the convolution integration electroencephalogram feature, the electrocardio data feature and the blood oxygen content of the newborn brain to generate the brain function feature.
Specifically, a two-dimensional convolution integration electroencephalogram feature map is converted into a one-dimensional feature map, then features of electroencephalogram data and features of cerebral blood oxygen data are fused, and a calculation formula is as follows:
in the above formula, the first and second carbon atoms are,
the characteristics of the brain function are shown,
the convolution is expressed to integrate the characteristics of the brain electricity,
representing the blood oxygen content of the newborn brain,
representing the electrocardiographic data.
S5023, the brain function state generating unit inputs the brain function features into the full-connection layer for classification, a convolutional neural network is used for generating the brain function state of the neonate, and setting parameters of the sub-low temperature bed and the hyperbaric oxygen chamber are regulated and controlled based on the brain function state of the neonate.
Specifically, the brain function features are input into the full-link layer for classification, and the calculation formula is as follows:
in the above formula, ωlA weight coefficient representing a full connection layer; blRepresenting the bias of the fully connected layer.
Specifically, the brain function state of the newborn comprises: dangerous state, severe ischemia-hypoxia brain function damage, mild ischemia-hypoxia brain damage and normal brain function state.
Identifying the functional state of the newborn brain by using a convolutional neural network, wherein the identification results are a, b, c and d; if the identification result is a, the newborn brain function state generating unit automatically and immediately calls medical care personnel corresponding to the identification result in a dangerous state; if the identification result is b, the corresponding identification result is that the ischemic and hypoxic brain injury is serious, and the newborn brain function state generating unit automatically changes the setting parameters (such as pressure, oxygen concentration, sub-low temperature setting temperature and the like) of the sub-low temperature bed and the hyperbaric oxygen chamber into a parameter setting scheme b when the brain injury is serious; if the identification result is c, the corresponding identification result is that the ischemic and hypoxic brain injury is light, and the newborn brain function state generating unit automatically changes the setting parameters (such as pressure, oxygen concentration, sub-low temperature setting temperature and the like) of the sub-low temperature bed and the hyperbaric oxygen chamber into a parameter setting scheme c when the brain injury is light; if the recognition result is d, the corresponding recognition result indicates that the brain function state is normal, and the newborn brain function state generating unit automatically changes the setting parameters (such as pressure, oxygen concentration, sub-low temperature setting temperature and the like) of the sub-low temperature bed and the hyperbaric oxygen chamber into a parameter setting scheme d when the brain function state is normal.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.