DTracker Artificial Intelligence based Diabetes Management Device

Project Title

DTracker Artificial Intelligence based Diabetes Management Device

Project Area of Specialization Artificial IntelligenceProject Summary Project Summary: The proposed system is an integrated version that includes smart devices and processors with special features of Information Processing, Artificial Intelligence (AI) and Instruction displaying on the output screen in order to aware the patient regarding his diabetes status. The working procedure includes daily monitoring of patient’s food intake and other health parameters which are related to diabetes. This monitoring eliminates the necessity of regular checkups to the endocrinologist. Monitoring of regular food intake of diabetic patient includes an AI algorithm developed to intelligently detect the calorie consumption of the patient. In addition, the health parameters measuring device integrated with the system can detect the number of calories burnt, level of blood pressure and the rate of pulse and respiration. As a result, it will maintain the health of the diabetic patient by monitoring whether the parameters have exceeded a threshold value or not. Therefore, this device will additionally benefit in accurate glycemic control, reduction in hypo-glycaemia and advancement in better quality of life. Project Objectives Objectives include:
•    SMART detection of type of food & calorie intake using Raspberry Pi integrated with Pi camera by applying an intelligent algorithm and training the dataset.
•    Designing of a gadget integrated with Pi will have an ability to evaluate the health parameters associated with diabetes such as the number of calories burnt by pedometer and other health measures (Test results of FPG and HbA1C, Blood Pressure, Pulse Saturation of Oxygen which provides Heartbeat and Respiration rate).
•    Designing of mobile application for maintaining the data of health parameters and calories and for tracking, predicting and providing the cautions regarding the health status. Project Implementation Method TASK 1: SMART DETECTION OF CALORIES AND FOOD TYPE:

• Dataset designing by collecting and capturing different food images.
• Data training by applying intelligent algorithm to detect the number of calories and food type from the user input.
• Inputting the patient captured image for Image Processing.
• Processed image is intelligently compared with the stored dataset to obtain the containing calories and the type of food.
• Setting the range of calories that can be consumed by the patient with respect to days.
• Investigating the date of the current day.
• If the date of the day has changed: 1. Current calorie dataset is set aside and stored to the mobile application. 2. Detected calories are added to a new calorie data.
• If the date of the day is same: Detected calories are added to the current calorie dataset.
• Checking whether the calories are in specified range:
• If the calories are in range: Let the patient eat. No warnings.
• If the calories have exceeded: Alarming condition. Display the warnings.

TASK 2 - DETECTION OF HEALTH MEASURES:

• Designing of a gadget that will be able to detect the no. of steps taken by the patient and other health measures that affects diabetic patients.
• If the health measures are not in range: Alarming condition. Display the warnings.
• Number of steps will determine the burnt calories. Furthermore, the burnt calories will be deducted from the consumed calories by the mobile app.
• All of these parameters will be stored on the application.

TASK 3 – DESIGNING OF MOBILE APPLICATION:

• Mobile application will be designed for user interface and to produce results by considering the information received from the hardware.
• At first, the ranges of all parameters will be defined to the application.
• Every information (regarding calories, health measures etc.) will be sent and processed by the app.
• The app will check whether the parameters are in range or not. 
• In addition, results of HbA1C test will send to the mobile application as well.
• It will notify regarding the alarming and in range conditions.
• After the completion of a month, it will provide precautionary measures and a predicted target to achieve in the upcoming month.

  Benefits of the Project According to the country profile of Pakistan 2019 provided by the International Diabetes Federation (IDF), over 19 million adults in Pakistan are estimated to be living with diabetes and putting them at risk of life-threatening complications. In the 9th Edition of the IDF Diabetes Atlas, it was published that Pakistan is in the top 10 countries for absolute increase in diabetes prevalence. Further, according to an article of National Center of Biotechnology Information (NCBI), all parts of the country have been affected, with the highest in Sindh and lowest in Khyber Pakhtunkhwa. 
This device is valuable for every diabetic patient as they are particularly vulnerable to severe and potentially life-threatening conditions. Regulation of this disease now necessitates the need of continuous observation and monitoring of health parameters as the increase in this disease is due to no control over the diet, increasing prevalence of obesity and physical inactivity. It will benefit the diabetic patient by detecting the health status and alarming conditions. Continuous monitoring of health measures leads to prediction of future complications of diabetes. It will be produced as a marketable & highly preferred product due to its high demand in such a crucial diabetic situation. Therefore, it will benefit the industries as well. Technical Details of Final Deliverable AI algorithms involves highly technical and specialized knowledge, this has not prevented AI from becoming an essential part of the medical technology and making contributions to major advances within the field of Biomedical Engineering. In D-Tracker, we use Raspberry-Pi as processing system interfaced with Pi Camera for detection of calories consumed by the diabetic patient. These calories will be detected with Machine Learning algorithm which is a major pillar of Artificial Intelligence and potential of AI to enable solutions has been investigated in the context of multiple critical management issues. Here we have used the AI algorithm to train the dataset in order to measure the food consumption which mainly majors the status of the patients’s health. Pi-camera will be interfaced which captures the image of the food and sends it to the Pi for Image Processed Raspberry Pi. Health parameter sensors such as Pedometer and other parameters detectors like Bp and SPO2 and will be integrated with the Raspberry Pi as well. Pedomerter will help in detecting the number of steps and basically the calories burnt by the patient. Outcomes of health measures such as Pulse Saturation of oxygen (Spo2) which involves Heart rate and Respiratory rate are necessary for further vital signs of the patient’s health. Their readings will be inputted to the Raspberry Pi for providing information to the device. Results of HbA1C test will also be fed to the system. Lastly, a system with an image processing technique using pi camera and an intelligent algorithm processor in addition to integrated pedometer and other health measure detectors will be proposed. This hardware information will be managed through a software (Mobile Application) which will provide a user interfacing environment. This Application will manage the data, track the health and cautions of the patient & predict the future health status.. Final Deliverable of the Project HW/SW integrated systemCore Industry HealthOther Industries Medical , Manufacturing Core Technology Artificial Intelligence(AI)Other Technologies Wearables and Implantables, Big DataSustainable Development Goals Good Health and Well-Being for People, Responsible Consumption and ProductionRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	78000
Raspberry Pi	Equipment	2	13500	27000
Arduino NANO	Equipment	2	4000	8000
Zigbee Module	Equipment	2	8250	16500
Pi Camera	Equipment	1	5000	5000
Pi Keyboard	Equipment	1	11500	11500
Miscellaneous	Miscellaneous	1	10000	10000