Open Access
Subscription Access
VITAVISION: Hybrid DenseNet-ResNet Architecture for Automated Vitamin Deficiency Detection with Explainability and Scalability
Abstract
Early detection of vitamin deficiencies is crucial for preventing health complications and ensuring well-being. This paper presents VitaVision, a cloud-based AI system that utilizes deep learning techniques to analyze images of the eyes, nails, and skin for detecting vitamin deficiencies. The system is trained on a curated dataset, employing an enhanced Convolutional Neural Network (CNN) model integrated with Multi-Head Self- Attention (MHSA) and an additional CNN module to improve accuracy and feature extraction. The proposed model is evaluated against standard architectures, including ResNet, DenseNet, and InceptionV3, using confusion matrices, accuracy scores, and loss curves. The results indicate that VitaVision outperforms conven- tional models in terms of classification accuracy and convergence speed. Additionally, the system provides personalized dietary recommendations based on the detected deficiencies, assisting users in addressing nutritional gaps effectively. The cloud-based deployment ensures accessibility for both general users and medical professionals, bridging the gap between AI-driven di- agnostics and healthcare accessibility. The system’s performance is validated through extensive experimentation, demonstrating its robustness and reliability in real-world scenarios. Future work includes dataset expansion, integration of real-time user data, and further optimization of the recommendation system to enhance accuracy and user experience.
References
1] D. A. Abuhani, J. Khan, and H. Sulieman, “Detecting Vitamin A Deficiency in Schoolchildren Using an Enhanced Explainable Machine Learning Model,” in Proc. Adv. Sci. Eng. Technol. Int. Conf. (ASET), 2023, pp. 1–5.
T. K. Chaithanya, J. Gupta, and M. S. Roobini, “Vitamin Deficiency De- tection Using Neural Networks,” in Proc. Int. Conf. Wireless Commun. Signal Process. Netw. (WiSPNET), 2024, pp. 1–10.
M. F. Ul Islam, M. Hasan, M. T. Rahman, and A. Chakrabarty, “Vitamin D Deficiency Detection: A Novel Ensemble Approach with Interpretability Insights,” in Proc. Int. Conf. Electr. Eng. Inf. Commun. Technol. (ICEE-[4] E. K. and S. K., “Diagnosis of Vitamin Deficiency in Human Beings Using DNN Algorithm,” in Proc. Int. Conf. Electron. Renew. Syst. (ICEARS), 2023, pp. 1627–1632.
E. K. and S. K., “Diagnosis of Vitamin Deficiency in Human Beings Using DNN Algorithm,” in Proc. Int. Conf. Electron. Renew. Syst. Based Med. Syst. (CBMS), 2024, pp. 51–56.
A. Khattak et al., “Proposing an Effective Deep Learning Model for Vitamin D Deficiency Diagnosis,” in Proc. IEEE Int. Symp. Comput.- Based Med. Syst. (CBMS), 2024, pp. 51–56.
J. Miryabbelli, “Deep Learning Technique DenseNet Application to Predict Vitamin Deficiency,” Ind. Eng. J., vol. 53, pp. 118–129, June 2024.
N. N. Lakshmi, M. J. Reddy, K. H. Krishna, and S. S. Reddy, “Vi- tamin Deficiency and Food Recommendation System Using Machine Learning,” Int. J. Res. Appl. Sci. Eng. Technol. (IJRASET), 2023.
J. Ramesh et al., “Explainable Machine Learning for Vitamin A Defi- ciency Classification in Schoolchildren,” in Proc. IEEE-EMBS Int. Conf. Biomed. Health Informatics (BHI), 2022, pp. 1–4.
I. Uvaliyeva, S. Belginova, and D. Borozenets, “Analysis of the Ma- chine Learning Methods Effectiveness for Morphological Classification of Anemia,” in Proc. Int. Symp. Multidiscip. Stud. Innov. Technol. (ISMSIT), 2023, pp. 1–5. ICT), 2024, pp. 137–142.
B. VeerasekharReddy, V. N. Thatha, G. U. Kiran, S. K. Shareef, N.
V. RajaSekharReddy and Y. R. Raju, ”Diet Recommendation System for Human Health Using Machine Learning,” 2023 7th International Conference on Computation System and Information Technology for Sustainable Solutions (CSITSS), Bangalore, India, 2023, pp. 1-5, doi: 10.1109/CSITSS60515.2023.10334162.
Refbacks
- There are currently no refbacks.