This paper explores the use of deep learning-based computer vision for real-time monitoring of the flow in intravenous (IV) infusions. IV infusions are among the most common therapies in hospitalized patients and, given that both over-infusion and under-infusion can cause severe damages, monitoring the flow rate of the fluid being administered to patients is very important for their safety. The proposed system uses a camera to film the IV drip infusion kit and a deep learning-based algorithm to classify acquired frames into two different states: frames with a drop that has just begun to take shape and frames with a well-formed drop. The alternation of these two states is used to count drops and derive a measurement of the flow rate of the drip. The usage of a camera as sensing element makes the proposed system safe in medical environments and easier to be integrated into current health facilities. Experimental results are reported in the paper that confirm the accuracy of the system and its capability to produce real-time estimates. The proposed method can be therefore effectively adopted to implement IV infusion monitoring and control systems.

Deep Learning-Based Computer Vision for Real-Time Intravenous Drip Infusion Monitoring

Liquids or medications administered through intravenous systems (infusion) is necessary when a patient requires immediate medication or requires slowly and continuously administered medications. This method is considered adequate in certain situations; however, the liquids/medications administration of through infusion has a risk when the nurse is late in replacing intravenous fluid. In this study, the system was designed to monitor the fluid infusion level by utilizing a short text message system. This system will provide information when the infusion fluid level is at 50 ml, 20 ml, and 0 via SMS. Moreover, a buzzer is utilized that acts as an alarm when the reaching 20 ml level, and infusion fluid have not been replaced yet, and when the intravenous fluid does not drip. Infra-red sensors and photodiodes are used to detect intravenous droplets of fluid, which are then used to calculate fluids' volume. The system is controlled using an Atmega 328 microcontroller. SIM Modem 900 is used to send SMS. Based on tests carried out, infusion fluid level monitoring systems have excellent performance. The percentage of system errors when detecting fluid level infuse is 1,21%. The function of sending fluid level information through SMS also works well

/pdf/short-text-message-based-infusion-fluid-level-monitoring-1excoegrk0.pdf

Short Text Message Based Infusion Fluid Level Monitoring System

The Artificial Intelligence (AI) is one of the emerging eras of at present. It supports big data analysis and the classification of them in fruitful way. Use of AI in real time data analytics provides optimized performance. Especially in the medical field, the AI plays important role. This research article focuses the kind of medical problem. The Intravenous Infusion System (IVF) has impressive role in transferring the medicine, saline and other liquids to the patients during hospitalization. The frequency of changing the IVF is taken care at most. So our proposed work discuss about designing the automated IVF system which has the three sections such as control, communicate and observe unit. This works on automated mode so whenever the saline low is below the predicted level the dripper unit triggers and adjusts the flow of saline with the patient's body temperature and other parameters variations. At the observer unit the clouds are connected. The monitoring data's of IVF changes are considered as base data set to train the machine learning models. This research also compares the performance matrices of the different machine learning methods such as accuracy, recall,F1 and precision. Here the SVM, NN, kNN & RF methods are compared where the RF has the better accuracy of 90.2%, SVM has 69.6 %, k NN has 69.5 % and the NN of general model has 92.5% for the obtained data set from the IVF device.

Performance Analysis of Smart Intravenous Infusion Systems using Machine Learning

Abstract In this paper, a method based on the usage of a smartphone for measuring simultaneously both heartbeat intervals and respiratory intervals is presented. In particular, the commodity accelerometer of a smartphone is used for measuring the seismocardiographic signal generated by heart activity and the acceleration due to breathing movements. The measurement is performed with the subject laying down, placing the smartphone on his/her xiphoid process. Signal processing algorithms are presented in the paper that can produce an accurate estimation of heartbeat and respiratory intervals from the measured acceleration signals. A metrological validation of the heartbeat and respiratory intervals estimates obtained with the proposed method is carried out by comparison with measurements obtained using an electrocardiograph and a spirometer. Two practical examples of applications of the measured quantities are finally reported, that are the measurement of the Heart Rate Variability from heartbeat intervals and of the Respiratory Sinus Arrhythmia from both the heartbeat intervals and the respiratory signal. 

Accurate simultaneous measurement of heartbeat and respiratory intervals using a smartphone

Additive manufacturing technologies allow the fabrication of smart objects, which are made up of a dielectric part and an embedded sensor able to give real-time feedback to the final user. This research presents the characterization of a low-cost 3D-printed strain sensor, fabricated using material extrusion (MeX) technology by using a conductive material composed of a polylactic acid (PLA)-based matrix doped with carbon black and carbon nanotubes (CNT), thus making the plastic conductive. A suitable measurement set-up was developed to perform automatic characterization tests using a high repeatability industrial robot to define either displacement or force profiles. The correlation between the applied stimulus and the variation of the electrical resistance of the 3D-printed sensor was evaluated, and an approach was developed to compensate for the effect of temperature. Results show that temperature and hysteresis affect repeatability; nevertheless, the sensor accurately detects impulse forces ranging from 10 g to 50 g. The sensor showed high linearity and exhibited a sensitivity of 0.077 Ω g−1 and 12.54 Ω mm−1 in the force and displacement range of 114 g and 0.7 mm, respectively, making them promising due to their low cost, ease of fabrication, and possible integration into more complex devices in a single-step fabrication cycle.

/pdf/additive-manufacturing-for-sensors-piezoresistive-strain-kfpwxvsq.pdf

Additive Manufacturing for Sensors: Piezoresistive Strain Gauge with Temperature Compensation

Intravenous (IV) infusion is one of the most common therapies in hospitalized patients. Monitoring the flow rate of the fluid that is being administered to the patient is therefore very important for his safety, considering that both over-infusion and under-infusion can cause serious health problems. In this document, a novel method for monitoring the flow rate in IV infusions is presented, that is based on deep learning computer vision techniques. Basically, the drip chamber is filmed with a camera and object detection is used to count drops. The proposed method is therefore less invasive than other ones developed for this purpose. Experimental results show that it can produce an accurate real-time estimate of the instantaneous flow rate of the drip. For these reasons, the proposed method can be effectively adopted to implement monitoring and control systems for health facilities.

Real-time drip infusion monitoring through a computer vision system

The increasing need to monitor Heart Rate (HR) has led to the development of dedicated devices based on miniaturized sensors and to their consequent pervasiveness. In this paper the aim is to develop a DAQ based system supervised by a LabVIEW Virtual Instrument to estimate the HR by means of an accelerometric plethysmograph. Two different signal processing techniques have been developed to estimate the HR respectively in time- and frequency domain, to study and compare their performances and to monitor the patient's heartbeat at rest by identifying possible pathological conditions or diseases.

Characterization of Heart Rate Estimation Using Piezoelectric Plethysmography in Time- and Frequency-domain

In this paper, a method to estimate the position and the entity of capacitive faults in coaxial cables based on the time domain reflectometry (TDR) is presented. A convolutional neural network (CNN) is used to analyze the reflectometric signals obtained from transmission lines containing multiple capacitive faults. The great quantity of data necessary for training the neural network was generated using a transmission line simulator. After the training procedure, the CNN was tested on both simulated and measured signals. The testing results prove that the neural network is capable to produce good estimates of the line characteristics, even when working with complex reflectometric signals.

Analysis of TDR Signals with Convolutional Neural Networks

This paper explores the use of a common smartphone for measuring simultaneously both heartbeat intervals and respiratory cycles. The proposed technique uses the smartphone’s accelerometer to measure the seismocardiographic signal and the acceleration due to breathing movements. The measurement is carried out while the subject is laying down, with the smartphone placed on his/her xiphoid process. In the paper, processing algorithms are presented, that can be used to obtain the heartbeat and the respiratory intervals from the measured signals. As concrete examples of possible application, heartbeat intervals are used to derive Heart Rate Variability and, together with the respiratory signal, to derive a Respiratory Sinus Arrhythmia measure on eight healthy volunteers.

Simultaneous Measurement of Heartbeat Intervals and Respiratory Signal using a Smartphone

The exploitation of Additive Manufacturing (AM) or 3D printing in the field of the sensor fabrication is abruptly rose in recent years. In this paper, new sensors manufactured using Fused Filament Fabrication (FFF) technology and two different commercial conductive filaments have been characterized from a thermal point of view. Using a climatic chamber to vary the temperature in a range between −5 °C and 50 °C, several tests were carried out showing a strong link between resistance and temperature changes. In particular, a positive correlation between resistance and temperature occurs. Different aspects due to the nature of the conductive 3D printed materials, nonconductive support structure, and the role of the air have been outlined, proving the need for subsequent thermal tests in order to exploit these low-cost devices as temperature sensors in the future.

Marco Scarpetta

Papers

Real-time drip infusion monitoring through a computer vision system

Characterization of Heart Rate Estimation Using Piezoelectric Plethysmography in Time- and Frequency-domain

Analysis of TDR Signals with Convolutional Neural Networks

Simultaneous Measurement of Heartbeat Intervals and Respiratory Signal using a Smartphone

Thermal Characterization of Electrical Resistance of 3D printed sensors