AI-POWERED, PORTABLE, AND AUTOMATED CARBOHYDRATE ANALYSIS AND INSULIN DOSE CALCULATION SYSTEM FOR DIABETES MANAGEMENT.
A portable, AI-powered diabetes management system automates carbohydrate analysis and insulin dose calculation using local AI and hardware-level processing, addressing user errors and ensuring continuous, accurate insulin delivery for improved treatment adherence and glycemic control.
Patent Information
- Application Number
- TR202522631
- Authority / Receiving Office
- TR · TR
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2025-12-29
- Publication Date
- 2026-10-21
AI Technical Summary
Current diabetes management systems, particularly for Type 1 diabetes patients requiring insulin supplementation, face challenges such as user-induced errors, intensive manual processing, high cognitive load, lack of analysis for homemade meals, and unsustainable accessibility, leading to potential complications and reduced treatment adherence.
A portable, AI-powered system that automates carbohydrate analysis and insulin dose calculation using a camera module for image processing, local AI, and hardware-level calculations, eliminating the need for external servers and ensuring data privacy, while being battery-operated for continuous use.
The system reduces calculation errors, improves treatment adherence, enhances glycemic control, and ensures uninterrupted diabetes management by providing real-time, accurate insulin doses without network dependence, suitable for various food types and environments.
Smart Images

Figure 00000013_0000
Abstract
Description
1 TARIFF AI-POWERED, PORTABLE, AND DIABETES MANAGEMENT SYSTEM. AUTOMATIC CARBOHYDRATE ANALYSIS AND INSULIN DOSE CALCULATION SYSTEM 5 Technical Area The invention relates to the monitoring and precision of regular blood glucose levels in conditions such as Type 1, Type 2, and gestational diabetes. In the management of chronic diseases requiring insulin dosing, manual carbohydrate 10 the risk of errors arising from counting and dose calculation processes and the user Developed to prevent cognitive strain; traditional nutritional chart by automating reading and complex mathematical processing steps Improves compliance and quality of glycemic control; tailored to meal contents and individual health. 15 digital images of meals via the camera module located within it capturing and analyzing these images using artificial intelligence-based image processing technology. by automatically determining food types and total carbohydrate amounts per serving identifying and analyzing processes as such; not outsourcing these processes to any external server or API service. by running it entirely on local hardware without needing to rely on internet 20 Keys on the system that ensure independence and the highest level of data privacy. The instantaneous blood sugar data received from the user through the team is analyzed. in accordance with standardized medicinal formulas, associating them with carbohydrate information. an analysis that automatically calculates the bolus insulin dose at the hardware level. The results and calculated dose are displayed to the user in real-time via the screen unit. 25 It offers the ability to operate without being connected to an external power source. Offering a portable design with an internal battery and power management unit; suitable for continuous use in the future. potential for integration with glucose meters and smart insulin pens It has an AI-powered automated carbohydrate analysis and insulin dosage system. It is related to the calculation system. 30 2 State of the Art Diabetes is a condition in which the body produces insufficient amounts of the hormone insulin or cannot use it effectively. a chronic metabolic disorder characterized by elevated blood sugar levels It is a disorder. This disease, which affects millions of people worldwide, has 5 in the management of Type 1 diabetes patients, especially those who require insulin supplementation, and insulin Regular blood sugar monitoring for patients with type 2 and gestational diabetes who exhibit diabetes resistance. And correct insulin dosing is vital. Today, the most commonly used nutritional strategy in diabetes management is 10 This is the "Carbohydrate Counting" method. This method involves counting the carbohydrates in the foods consumed. Bolus insulin to be administered according to the amount and instantaneous blood sugar level. It enables the determination of the dose. Medical advice for calculating insulin dose (ID). A standardized mathematical formula is used in the literature: ID = (TKH / 15) + ((BG - 100) / 50) The variables TKH (Total Carbohydrate) and DS (Blood Sugar) in this formula while varying for each meal; coefficients used in division operations (Values 15 and 50 in the example) refer to the patient's insulin sensitivity and carbohydrate / insulin ratio. These are parameters determined individually by physicians according to the ratio. The coefficients can be adjusted differently for each patient. In the current state of the art, this calculation process is largely manual or semi-manual. It is carried out using automated methods: 25 • Manual Calculation Method: Patients are calculated by healthcare professionals. Using the standard nutritional tables provided, the nutrients in each meal It determines carbohydrate values in grams or spoons and manually. It sums them up as follows. Then, the complex mathematical 30 mentioned above... by performing the procedures themselves, they can determine the required insulin dose. They determine. • Nutritional Database Applications: Offering a comprehensive nutritional database. However, these applications are not directly a carbohydrate counting module. 3 It does not contain any. Users must search for each nutrient individually and note their values. and the calculation must be done manually. • Digital Carbohydrate Counting Systems: For direct carbohydrate counting. The designed systems, however, require the user to manually enter a large amount of data for each meal. These systems demand pre-prepared foods. 5 their inclusion, their inadequacy in the analysis of traditional and homemade dishes This is the reason. The technical limitations and disadvantages of the current methods are as follows: • Risk of User-Induced Errors: The calculation process may involve the user's mathematical knowledge. it depends on skill, education level and attention, incorrect dosage This can lead to adverse effects; this situation disrupts blood sugar regulation. This increases the risk of complications in the long term. • Intensive Manual Processing, Time Loss, and High Cognitive Load: At every meal and Manual procedures that need to be repeated between meals, time spent on patients 15 It creates a burdensome and exhausting load, negatively affecting treatment adherence. • Lack of Analysis: Manual methods and limited digital databases, visual They are unable to determine portion sizes, especially when preparing complex homemade meals. They are unable to determine the correct amount of carbohydrates in their meals. • Sustainability: Especially for older individuals, those with limited mathematical skills (20 years old) for patients who are illiterate or unable to read and write, the available methods It is very difficult to implement and is not sustainable. • Uninterrupted Accessibility and Internet Independence: The entire artificial intelligence of the system By performing its operations on its own processor, the internet 25 Even under challenging environmental conditions, the diabetes management process continues uninterrupted. It is sustainable. • Maximum Data Privacy: Personal health data and privacy information Not transferring meal images to any external network reduces the risk of data leakage. eliminating and protecting biometric data security across device integrity 30 It is included. 4 • Low Latency: Data is sent to and processed by remote servers. By eliminating network-related delays in the process, analysis and dose calculation The results are presented to the user in real time. Given all these limitations and technical shortcomings, diabetes 5 the laborious nature of manual calculation processes and intensive data entry in management This need negatively impacts patient adherence to treatment, leading to serious health problems. This carries the risk of complications, especially with standard insulin doses. included in the calculation formula and according to medical requirements for each patient. 10 with a carbohydrate / insulin ratio that needs to be precisely adjusted by experts. critical parameters such as insulin sensitivity factor (examples 15 and 50 in the formula) flexible and error-free in existing manual or semi-automatic systems (such as their values) Managing it in this way is technically quite difficult. This technical and operational aspect... bottlenecks; by automatically analyzing meal images with artificial intelligence algorithms 15 personalized parametric data that determine the carbohydrate values of foods It operates within its system and determines the automatic insulin dose at the hardware level. by offering this entire process in a portable format without relying on an external device reducing the margin of calculation error and providing sustainable diabetes management. This makes the need for an integrated system inevitable. In conclusion, due to the negative aspects described above and the existing solutions... Due to the lack of knowledge on the subject, an improvement should be made in the relevant technical field. It was deemed necessary. Purpose of the Invention 25 The invention introduces a new approach to this field, different from the structures currently used in the technology. the creation of a structure with different technical features that brings about innovation It aims to... The primary aim of the invention is to eliminate the disadvantages detailed above, vital in diabetes management, but high when performed manually. Carbohydrate counting and bolus insulin dosage create a risk of error and cognitive load. computational processes, AI-powered image processing technology and integrated A portable system that fully automates through hardware components. to present. Another purpose of the invention is to capture photos of meals via the camera module. Using artificial intelligence technology to analyze, categorize food types and portion-based 5 determines the total carbohydrate amounts, then enters them via the keypad. By combining this analysis with instantaneous blood sugar data, an accurate and reliable dosage result is obtained. The goal is to create a technical system that produces results. Another purpose of the invention is to create a single-board computer, camera, screen, and keyboard, among other things. by combining the components with a battery unit that has an intelligent power management circuit, each The goal is to design a portable hardware architecture that can be used in various locations. Another purpose of the invention is to adjust the coefficients included in the dose calculation formula. (carbohydrate / insulin ratio and sensitivity factor) 15 on a user-by-user basis allowing for adjustment, reducing manual calculation errors, and extending the lifespan. It helps prevent long-term diabetes complications and improves health. The goal is to develop an innovative system that supports sustainability in technologies. Another aim of the invention is to differentiate between the age, education level, and mathematics (20) of individuals with diabetes. regardless of their skills, instead of complex data entry intuitive and accessible, working with just taking photos and simple value input. The goal is to develop a user interface architecture. This will, in particular, benefit older individuals. or sustainable diabetes management for patients who have difficulty with calculations. The aim is to provide a system that improves treatment adherence and compliance. 25 Another aim of the invention is to improve the insulin dose determination process at the hardware level. by standardizing, eliminating user-generated calculation errors and personal judgments This creates a highly secure system that minimizes the possibility of incorrect dosages. to place. Another aim of the invention is to improve glycemic control and prevent long-term complications. a technical system aimed at reducing risks and frequency of hospitalizations It is to design. 6 Another aim of the invention is to use artificial intelligence technology to process not only ready-made foods. not only, but also including traditional and homemade dishes, it is constantly updated and a system that performs nutritional analysis through a constantly evolving pool of information The goal is to develop this technological infrastructure and produce almost limitless varieties of high-quality food. 5 that can identify with a high degree of accuracy and increase the precision of the analysis over time The goal is to provide a universal analysis system. Another purpose of the invention is to record daily dietary habits and blood sugar trends. automatically stores and tracks insulin usage patterns and this critical data. Digital 10, which allows for safe sharing with healthcare professionals. The goal is to design a healthcare infrastructure that facilitates communication between patients and doctors. empowering, offering decision support mechanisms based on real-time data. The goal is to develop an integrated system. Another aim of the invention is to create real-time blood glucose monitors and 15 other devices in the future. By enabling wireless integration with smart sensor technologies, it provides instant blood sugar measurements. data directly from the relevant devices without requiring user intervention. The goal is to provide an automated system. With this technological expansion capacity. User-generated data entry that makes the data entry process completely autonomous. a fully integrated digital ecosystem in diabetes management that eliminates errors 20 to reveal. Another purpose of the invention is to prevent acute illnesses resulting from erroneous dose calculations. by minimizing complications, the long-term cost to the healthcare system The aim is to offer a system that reduces the burden and saves on treatment costs. 25 The structural and characteristic features and all the advantages of the invention are given below. through detailed explanations written with figures and references to these figures This will be understood more clearly, and therefore the evaluation will also be based on these methods and... This must be done taking into account the detailed explanation. 30 7 Figures that will help understand the invention. Figure 1 shows an AI-powered, portable and automated system for diabetes management. with all subcomponents of the carbohydrate analysis and insulin dose calculation system It is a joint representation. 5 Drawings do not necessarily need to be scaled and are useful for understanding the existing invention. Details that are not necessary may have been omitted. Furthermore, the most important... functions that are at least substantially identical or at least substantially identical The elements that exist are indicated by the same number. 10 Explanations of Figure References 1. AI-Powered, Portable and Automated Carbohydrate System for Diabetes Management Analysis and Insulin Dose Calculation System 15 1.1. Central Processing Unit (Single Board Computer) 1.2. Image Capture Unit (Camera Module) 1.3. Display Unit (Screen Module) 1.4. Data Input Unit (Keypad Module) 1.5. Power Management Unit (Battery Module) 20 Detailed Description of the Invention In this detailed explanation, the preferred configurations of the invention are presented, not only in terms of the subject matter but also in a more comprehensive way. 25 It is explained. The invention is critical for carbohydrate counting and insulin management in diabetes administration. by minimizing dose calculation processes, user intervention, and margin of error It is a system that automates. 30 As shown in Figure 1, the system (1) basically consists of a central processing unit (1.1), This unit includes an image capture unit (1.2) and a data input unit (1.4), which transmit data to this unit. an imaging unit (1.3) that presents the results and a power supply that feeds the entire system. It consists of management units (1.5). 35 8 The Central Processing Unit (1.1) is the main decision and processing center of the system. Image Built-in capture unit (1.2) to process data coming from it. It incorporates an artificial intelligence model. It does not require access to any external server or API. Without needing a separate serving; food identification, portion analysis, carbohydrate counting 5 and carries out its insulin dose calculation processes entirely with local resources. The Image Capture Unit (1.2.) captures a digital image of the meal to be analyzed. optical captures and transmits to the central processing unit (1.1) for local AI analysis. It is an input unit. 10 The Imaging Unit (1.3) displays nutrients detected by the local artificial intelligence model. values and automatic insulin calculated by the central processing unit (1.1) It is the interface that presents the dosage result to the user. The Data Entry Unit (1.4.) records the user's blood sugar value and personal treatment. entering the parameters into the system, and the system initiating local processing processes. It is the interaction unit that enables and confirms this. Power Management Unit (1.5.), artificial intelligence on central processing unit (1.1) 20 designed to meet the high processing power requirements of the model, It is the unit that ensures the portability and power continuity of the system. The working process of the system (1) will be analyzed by the image capture unit (1.2). It begins with taking a digital image of the meal. This captured image is then processed by the central processing unit 25 is transmitted to the central processing unit (1.1). The central processing unit (1.1) contains artificial intelligence. Object recognition and portion analysis on the image through the model. This analysis is carried out by identifying the types of food served at the meal. Total carbohydrate (TKH) per serving is automatically calculated in grams. is determined. The imaging unit (1.3) detects 30 by the central processing unit (1.1). It presents the nutritional information and total carbohydrate content for user approval. The user enters the instantaneous measured blood glucose (BG) value via the data input unit (1.4). enters the system (1). The central processing unit (1.1) inputs the total carbohydrate data obtained. and using the blood sugar data entered by the user, the following mathematical formula is used: 9 Calculates the bolus insulin dose (ID) according to the formula and displays the unit (1.3) It presents it to the user via: ID = (TKH / A) + ((KŞ - 100) / B) In this formula, A (carbohydrate / insulin ratio) and B (insulin sensitivity factor) are 5. Although the values are exemplified as 15 and 50 respectively in standard practice, the system (1) these coefficients are adjusted on a patient basis in line with physician recommendations It has a flexible software architecture that allows for these possibilities. These automated processes, run by the central processing unit (1.1), are traditional 10 mathematical errors frequently encountered in manual calculation methods and It eliminates nutritional information inaccuracies. The system (1) will, in the future, use continuous data entry via the data entry unit (1.4) instead of manual entry. Automatic blood glucose data from glucose meters via wireless protocols 15 It has the capacity for expansion, which it can take on as such.
Claims
REQUESTS 1. Manual carbohydrate intake in the management of Type 1, Type 2, and gestational diabetes. to prevent risks arising from counting and dose calculation errors and glycemic Developed to improve control quality, AI-powered, 5 A portable and automated carbohydrate analyzer and insulin dose calculator. The system (1) has the following features; - An image that captures the contents of the meal to be analyzed as digital data. capture unit (1.2), - User's real-time blood glucose (BG) data and personalized treatment 10 a data input unit (1.4) that allows the input of parameters, - Image analysis results and calculated insulin dose amount. a display unit that presents to the user (1.3), - The system can operate without being connected to an external power source. a power management unit (1.5) and 15 - Data from the image capture unit (1.2) to any external a built-in system that analyzes locally without sending it to the server by incorporating an artificial intelligence model, it identifies food types and total amounts based on portion sizes. Determines the amount of carbohydrates (TKH); this data is obtained from a user's blood sample. By correlating the bolus insulin dose with blood sugar data, hardware-level adjustment is possible. a central processing unit that calculates automatically (1.1) It includes.
2. The system mentioned in claim 1 is (1) and its feature is that the central processing unit (1.1), object recognition on data received via image capture unit (1.2) and by running portion size estimation algorithms locally 25 It is the process of determining the total carbohydrate (TKH) content of foods.
3. The system mentioned in claim 1 is (1) and its feature is that the central processing unit (1.1), the detected total carbohydrate (TKH) data and data input unit (1.4) Blood glucose (BG) data obtained from the sample is calculated as follows: BG = (TKH / A) + ((BG - 100) / B) It is the processing at the hardware level according to the mathematical formula. 30 4. The system mentioned in claim 3 is (1) and its feature is; A in the formula (carbohydrate / insulin ratio) and B (insulin sensitivity factor) coefficients, data via the input unit (1.4) for each user in accordance with physician recommendations The advantage is that they can be defined and updated individually. 11 5. The system mentioned in claim 1 is (1) and its feature is that the central processing unit (1.1), analyze the clarity of the image captured by the image capture unit (1.2). by doing so, the imaging unit (1.3) in cases unsuitable for analysis It is the process of sending a warning to the user.
6. The system mentioned in claim 1 is (1) and its feature is that the central processing unit (1.1), 5 from external glucose meters and / or smart insulin pens It will be able to automatically receive data via wireless communication protocols and a system that can include the user in the calculation process without any intervention It is about having technical capacity.
7. The system mentioned in claim 1 is (1) and its characteristic is that the central processing unit (1.1), 10 For long-term monitoring and reporting of system data, a mobile The application component will be able to interact with the data synchronization in the future. with the technical capacity, integrated software architecture and It has a wireless communication infrastructure.
8. The system referred to in Claim 1 is (1) and its feature is; the power management unit (1.5), 15 High processing power of the native AI model on the central processing unit (1.1) energy storage capacity and voltage to meet power requirements It includes a battery module with a regulating circuit.
9. A method that complies with Claim 1 and has the following characteristics: - Digital 20 of the meal to be analyzed via the image capture unit (1.2) obtaining the image, - The resulting image is located within the central processing unit (1.1). by the artificial intelligence model found, to any external server or food analyzed locally without being sent to an API service. Identifying types and portion sizes, 25 - Local analysis results show the total carbohydrate (TKH) amount in the meal. determined in grams - The specified nutrient content and total carbohydrate amount Presented for user approval via the display unit (1.3), - User's instantaneous blood glucose (BG) data via data entry unit (1.4) 30 and / or personalized treatment coefficients (A and B) into the system (1) definition, - By the central processing unit (1.1); total from local analysis Carbohydrate data and user blood glucose data, ID = (TKH / 12 A) + ((KŞ - 100) / B) processed according to the mathematical formula The bolus insulin dose (BID) that should be administered is automatically determined calculation, - Display unit of the calculated final insulin dose result (1.3) 5 presented to the user through It includes the steps of the process.
