Impact of Nutritional Factors on Blood Glucose Levels in Diabetic Patients

Abstract

Diabetes management requires continuous monitoring of blood glucose levels, where dietary intake plays a critical role in influencing glycemic variations. This paper presents an intelligent system that analyzes the impact of nutritional factors on blood glucose levels using food image recognition. The proposed approach employs a Convolutional Neural Network (CNN) to automatically identify food items from user-uploaded images. Nutritional information, particularly carbohydrate content, is then extracted from a predefined dataset and used to estimate the potential glucose impact. A rule-based classification method is applied to categorize glucose levels into low, normal, and high, providing an interpretable assessment of dietary effects. Furthermore, a recommendation module suggests suitable food alternatives to help maintain balanced glucose levels. The system is implemented using Python, Flask, and TensorFlow/Keras, ensuring a lightweight architecture with fast response time. Experimental results demonstrate that the system performs efficiently for known food categories and provides consistent predictions. The proposed solution assists diabetic patients in making informed dietary decisions and highlights the potential of integrating artificial intelligence with nutritional analysis for effective healthcare applications.

Introduction

The text describes diabetes mellitus as a chronic condition requiring careful blood glucose management, especially for Type-1 patients who depend on insulin therapy. Diet, particularly carbohydrate intake, plays a critical role in controlling blood glucose levels, but traditional methods of tracking food intake are often inaccurate and inconvenient.

To address this, the study proposes an AI-based system that uses deep learning (CNN) to identify food from images, estimate nutritional values (mainly carbohydrates), and predict glucose impact (low, normal, high). It also provides personalized dietary recommendations to help patients manage their condition effectively.

The literature review highlights that while machine learning models have improved glucose prediction, most systems rely on extensive data or monitoring devices. There is limited research on using food images directly, creating a gap that this system aims to fill.

The proposed methodology involves uploading a food image, preprocessing it, recognizing the food using CNN, retrieving nutritional data, and classifying glucose impact through a rule-based approach. Based on the results, the system generates recommendations and displays all information through a user-friendly interface, offering a fast and practical solution for diabetes management.

Conclusion

This paper presented a system for analyzing the impact of nutritional factors on blood glucose levels in diabetic patients using food image recognition. The proposed approach integrates a Convolutional Neural Network (CNN) for food classification with nutritional analysis and a rule-based method for glucose prediction. The system effectively identifies food items, estimates carbohydrate content, and is limited by dataset size and lacks personalization classifies glucose impact into low, normal, and high categories. The system effectively identifies food items from images, estimates their carbohydrate content, and classifies the glucose impact into low, normal, and high categories. The inclusion of a recommendation module further enhances the system by suggesting appropriate dietary options to help maintain balanced glucose levels. Experimental results demonstrate that the system provides consistent predictions with low response time and satisfactory performance for known food categories. The recommendation module further enhances the system by suggesting appropriate dietary options based on predicted glucose levels. Although the current implementation is limited by dataset size and lacks personalization, it provides a simple and efficient solution Future work will focus on improving model accuracy, incorporating personalized health data, and integrating real-time monitoring systems

References

[1] J. Contreras and J. Vehi, “Artificial intelligence for diabetes management and decision support,” Journal of Diabetes Science and Technology, vol. 12, no. 2, pp. 332–339, 2018. [2] S. Mougiakakou et al., “Artificial intelligence in diabetes management,” Computational and Structural Biotechnology Journal, vol. 16, pp. 158–168, 2018. [3] G. Ciocca, P. Napoletano, and R. Schettini, “Food recognition: A new dataset, experiments, and results,” IEEE Journal of Biomedical and Health Informatics, vol. 21, no. 3, pp. 588–598, 2017. [4] M. Chen, Y. Hao, K. Hwang, L. Wang, and L. Wang, “Disease prediction by machine learning over big data,” IEEE Access, vol. 5, pp. 8869–8879, 2017.[5] A. Alamri, M. M. Hassan, and A. Alelaiwi, “Applications of IoT in healthcare for diabetes management,” Future Generation Computer Systems, vol. 97, pp. 356–365, 2019. [5] N. Zhu, Y. Liu, and X. Zhao, “Food image recognition based on deep learning for dietary assessment,” IEEE Access, vol. 7, pp. 123000–123010, 2019. [6] J. Beutel et al., “Continuous glucose monitoring: Clinical applications and challenges,” Sensors, vol. 19, no. 3, pp. 1–15, 2019. [7] Z. Zhu, E. C. C. Chan, and B. Wang, “A deep learning approach for blood glucose prediction using CNN-LSTM model,” IEEE Access, vol. 8, pp. 123456–123465, 2020. [8] C. Dalla Man, R. A. Rizza, and C. Cobelli, “Meal simulation model of the glucose-insulin system,” IEEE Transactions on Biomedical Engineering, vol. 54, no. 10, pp. 1740–1749, 2007. [9] R. Hovorka, “Continuous glucose monitoring and closed-loop systems,” Diabetic Medicine, vol. 23, no. 1, pp. 1–12,2006

Copyright

Copyright © 2026 Mrs. Vasantha P, Nandini V, Mehreen Sulthana Sk, Apurupa Y. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.