AI-Powered Visual and Thermal Monitoring System for Tracking Child Health Parameters and Its Working Method

Abstract

The invention relates to an artificial intelligence-assisted monitoring system and operating method developed for the remote and continuous monitoring of vital health parameters of children and newborns without any physical contact. The invention is characterized by its sensor module containing at least one visual light spectrum (RGB) camera and at least one thermal camera that collects real-time data from the child, a communication unit that transfers this data, and a central processing unit that analyzes the data using multispectral sensor fusion and deep learning algorithms to extract respiratory rate, abdominal and chest movements, body temperature changes, and movement patterns.The system described in this invention has three different operational application modes: a local network architecture where data is processed on an external local computer over a local network, a cloud-based architecture where data is processed on a remote cloud server over a wide area network, and an integrated edge artificial intelligence architecture where data is processed locally on a GPU-based embedded processor directly on the device chassis without requiring any external network transmission. The system also includes a clinician monitoring and graphical interface that presents vital parameter trends live with color-coded triage levels, and a multi-level clinical alarm and notification mechanism that generates simultaneous bedside, central, and mobile emergency calls in critical situations.

Description

1 TARIFF 5 AI-POWERED VISUAL AND THERMAL MONITORING CHILD HEALTH PARAMETERS TRACKING SYSTEM AND WORKING METHOD Technological Field 10 This invention allows for the measurement of vital health parameters in children (respiratory rate, body temperature changes, and movement). (patterns) are tracked contactless using thermal cameras, RGB cameras, and artificial intelligence. processing the data, providing the resulting analysis results and risk assessments to the user / healthcare personnel. It is related to a tracking system and working method that transmits information. State of the Art 15 Current practices in the fields of child and infant care and medical monitoring technologies. They typically focus on limited analytical capabilities or operate with a single type of sensor. Traditional systems monitor the baby's environmental conditions or simply use standard RGB. features such as providing video streaming and / or audio-based alerts via cameras It is limited. 20 However, such solutions are insufficient for deeply analyzing the immediate health status of children. For example, standard cameras cannot collect clear data in dark environments. It is negatively affected by changes in light and microchest during respiration. It cannot accurately detect movements. Furthermore, it only uses electrodes that come into contact with the body, or... Measurements taken with wearable sensors indicate irritation, stress, and 25% sensitivity in children's delicate skin. This leads to a risk of infection. Many of the current systems used for image analysis are real-time, high-performance. It lacks an artificial intelligence support mechanism. Furthermore, it processes data using a local processor instead of... We are constantly forced to transfer data to external cloud servers, which poses data privacy risks. It also causes the system to become inoperable during network outages. 30 in this area technical limitations, more integrated, contactless, multi-sensor fusion-based and flexible hardware This requires the development of intelligent solutions with unique architectures. Description of the Invention This invention is an AI-powered visual system that can overcome the aforementioned disadvantages. and a system and working method for monitoring child health parameters through thermal monitoring, 35 Its feature is that it has at least one visual light spectrum (RGB) camera and at least one thermal camera together. It incorporates sensor fusion technology, and the data obtained is used in advanced artificial intelligence. Contactless analysis of respiration, instantaneous thermal changes and movement is performed by processing with algorithms. This can be done and the system can be either embedded on the device, on the local network, or depending on operational needs. It is also able to operate flexibly in cloud-based architectures. 40 Thanks to this invention, children can be examined in complete darkness without any sensors touching their bodies. Even in these environments, respiratory rate abnormalities, sudden temperature changes, and sleep / movement disturbances are observed. 2 It is detected with high accuracy. This integrated approach reduces the need for multiple medical devices. 5 by eliminating these obstacles, it saves costs and space while benefiting healthcare personnel and parents. It reduces stress. Explaining the Figures The invention will be described with reference to the attached figures:  Figure 1: This is a block diagram showing the general hardware components of the system. 10  Figure 2: System architecture with external local computer / server connection (LAN / WLAN) It is a method of application.  Figure 3: System architecture with remote cloud server over wide area network (WAN / Internet) It is a method of application.  Figure 4: Integrated embedded architecture where the main components of the system are grouped within the same body 15 It is a method of application. Explanation of References 1. Visual Light Spectrum (RGB) Camera 2. Thermal Camera 3. Sensor Module 20 4. Communication Unit 5. Central Processing Unit 6. Local Computer / Server 7. Cloud-Based Server 8. Embedded Processor 25 9. Device Body 10. Clinician Monitoring and Charting Interface 11. Multi-Level Clinical Alarm and Notification Mechanism Disclosure of the Invention The invention essentially involves at least one visual light that collects real-time data from the child. a sensor module (3) containing an RGB spectrum camera (1) and at least one thermal camera (2); a communication unit that enables the transfer of data (4) and an artificial intelligence-based system that processes the data It consists of a central processing unit (5) components with algorithms. The sensor module (3) in the system uses visual data from the child and infrared thermal data. It simultaneously captures radiation data. The communication unit (4) captures this multispectral data 35 It transmits data to the central processing unit (5) via Wi-Fi or Bluetooth depending on the system architecture. The central processing unit (5), through the artificial intelligence algorithms running on it; RGB camera (1) from the images the child’s physical position, facial-hand-arm movements and abdomen-chest When analyzing the movements, the child's face and body surface were analyzed from the thermal camera (2) data. It detects temperature changes, especially around the nostrils, during breathing. 40 The RGB camera may be insufficient for positioning in low light conditions. Thermal camera images are used in these situations. The artificial intelligence model uses these two different data points. By blending the source (sensor fusion), the error margin of the vital parameter is minimized. It produces results. These results are used to provide meaningful medical insights through clinician follow-up and graphical interface (10). are transformed into trends. The artificial intelligence algorithm analyzes the child's respiratory trends and sudden thermal changes. By analyzing fluctuations retrospectively, the system provides clinicians with graphical reports. It functions as a clinical decision support mechanism to lighten the clinician's workload; 3 Multi-level clinical alert and 5 when parameters return to normal or approach critical levels. through the notification mechanism (11) the clinician is notified with visual, auditory or remote mobile notifications This alerts the clinic, thereby optimizing the infrastructure for early diagnosis and rapid intervention. Detailed Description of the Invention The system described in this invention can be configured into three different types depending on the usage scenarios and infrastructure requirements. It can work with hardware and data transmission architecture: 10  Local Network Architecture (Figure 2): In this application configuration, the network houses the cameras (1, 2). The sensor module (3) transmits data to a local network (LAN / WLAN) via the communication unit (4) via, to a local computer physically located separately from the system, or It transfers it to the server (6). The central processing unit (5) is located inside this local computer (6). The analysis results obtained are used in the clinician follow-up and graphical interface (10) and clinical alarm and 15 It is reported to healthcare workers through the notification mechanism (11). Especially in-hospital children This architecture is preferred for installations that do not want data to leave the building during service deployments. It is done.  Cloud-Based Architecture (Figure 3): In this application, the sensor module (3) collects raw or pre-processed data communication unit (4) and a wide area network (WAN / Internet) 20 It transmits the data to a cloud-based server (7) located remotely via artificial intelligence. All of the analyses are on the central processing unit (5) of this remote cloud-based server (7) It is performed by. Similarly, clinician monitoring and graphical interface (10) and Users are informed via the clinical alarm and notification mechanism (11). Home type in uses, mobile application integrations and remotely monitored subscriptions 25 This architecture provides an advantage in their models.  Integrated Embedded AI Architecture (Figure 4): In this application type, external The need for a computer or internet connection has been completely eliminated. Centralized The processing unit (5); sensor module (3) together with the same device housing (9) or electronic a high-performance, GPU-based embedded 30 integrated into the card is a processor (8) (e.g., Jetson Nano or similar hardware). Artificial intelligence algorithms, The raw images are captured entirely locally and on the device itself, without exporting them. Time-bound tasks. The analyses obtained are used for clinician monitoring and graphical interface (10) and clinical alarm. and is notified to healthcare workers through the notification mechanism (11). This architecture maximizes Data privacy, zero network latency, and a portable, standalone medical monitor device form factor are achieved. 35 It enables it to be done. Production and Integration of the Invention: The physical production of the system involves industrial standard plastic injection molding or three-dimensional printing. a device housing (9) shaped by technologies, containing the components This is achieved by integrating electronic circuit boards (PCBs). On this structure, 40 circuit boards are placed relative to each other. calibrated visual light spectrum (RGB) camera (1) and thermal camera (2) components The sensor module (3) containing the data is positioned. According to the system architecture, the one that will process the data is located. central processing unit (5); on-board GPU-based embedded processor (8) communication unit (4) (Wi-Fi, Bluetooth, Ethernet or cellular network) can be placed as well. via modules) to an external local computer (6) or a remote cloud-based server (7) 45 It can be configured programmatically to connect. The developed deep learning, Computer vision and sensor fusion algorithms, software embedded in the relevant hardware, or The system becomes functional by being installed as a server-based application. 4 Applications of the Invention: 5 The system described in this invention addresses a wide range of medical and civil applications requiring the monitoring of child health. Suitable for use in the following environments:  Hospitals and Clinical Settings: Neonatal intensive care units, premature babies incubators, pediatric wards, pediatric emergency areas, and resuscitation rooms as a bedside monitor or central monitoring system, 10  Home and Smart Living Spaces: In baby and children's rooms, parents can use their children's rooms to... sleep quality, respiratory status, movement abnormalities, and instantaneous body temperature. a smart home health monitor that can track changes remotely and seamlessly aspect,  Nurseries, Daycare Centers and Childcare Centers: Places where many children are together 15 institutional settings where health, fever, and movement monitoring should be carried out collectively. It can be used as a central early warning and security infrastructure in these areas.

Claims

REQUIREMENTS 5 1. The invention is a system for monitoring child health parameters, the characteristic of which is;  at least one visual light spectrum (RGB) that collects real-time visual data from the child camera (1),  at least one thermal camera collecting simultaneous thermal data from the child (2),  a sensor containing the RGB camera (1) and thermal camera (2) components 10 module (3),  a communication unit (4) that transmits data from the sensor module (3) and  By processing this data with artificial intelligence-based algorithms, we can measure the child's respiratory rate and body temperature. a central processing unit that analyzes temperature change and motion patterns (5) It includes. 15 2. According to Claim 1, it is a monitoring system for child health parameters, characterized by its centralized processing. a local computer located physically separate from the unit's (5) sensor module (3) or server (6) and communication unit (4) transmits the said data over a local network It is structured to transfer to a local computer or server (6).

3. A monitoring system for child health parameters according to Claim 1 or 2, characterized by its central location. 20 a cloud-based server located at a distance from the sensor module (3) of the processing unit (5) (7) and the data of the communication unit (4) through a wide area network (Internet) to this cloud It is structured to transmit to the server (7).

4. According to Claim 1, it is a monitoring system for child health parameters, characterized by its central processing capability. unit (5) is integrated on the same device housing (9) as the sensor module (3) 25 Having a GPU-based embedded processor (8) and transmitting artificial intelligence analyses via an external network It has a hardware structure that performs this locally without needing to be physically present.

5. Monitoring system for child health parameters according to any of claims 1, 2, 3 or 4. its feature is the respiration, temperature and motion analysis produced by the central processing unit (5). A clinician's tracking and graphing system that displays results in real-time by converting them into time series graphs. 30 the interface (10) and the risk status of the child in the said interface (10) beforehand It will display clinical triage levels live using color-coded settings based on defined threshold values. It is in the structure.

6. The invention is a working method for a monitoring system of child health parameters, characterized by:  Simultaneous visual 35 through the child via the visual light spectrum (RGB) camera (1) data acquisition,  Simultaneous infrared thermal radiation through the child via thermal camera (2) data collection,  The visual data and thermal data received are processed centrally via the communication unit (4). transfer to unit (5), 40  Artificial intelligence algorithms running in the central processing unit (5) from visual data Determining the child's physical position and abdominal-chest movements,  Thermal data reveal the child's instantaneous temperature changes and changes in breathing. analysis of thermal differences and  By blending the detected visual and thermal features with sensor fusion, respiratory 45 generating vital parameter results such as speed, body temperature change and movement It consists of process steps. 6 7. The working method of the system in accordance with claim 6, and its feature is; in the central processing unit (5) 5 If the calculated vital parameters are found to be at a clinically critical level, activation of the multi-level clinical alarm and notification mechanism (11), bedside Simultaneous generation of audible and visual alarm signals on the monitor and communication unit (4) via the responsible clinician's mobile device or hospital central nurse call It includes the steps involved in sending an immediate emergency notification to the system. 10