AI-Powered Real-Time Medical Diagnosis for Low-Resource Settings: A CNN-Based Edge AI and Blockchain-Integrated Solution V. Vijaya Rama Raju, I.Ranaveer, Haider Mohmmed Alabdeli, P.Selvaprasanth, Kapesh Subhash Raghatate, Leoni Sharmila S 2025 International Conference on Metaverse and Current Trends in Computing Icmctc 2025, 2025 In low-resource settings, healthcare use and access are still very much limited by a lack of infrastructure, shortage of medical expertise, and medical connectivity. Given that, this paper presents a novel framework to have an AI-powered real-time diagnosis of medical problems we have with real Convolutional Neural Networks (CNNs), edge AI, and federated learning to customize accurate and scalable healthcare in remote areas. The proposed system consists of integrating lightweight CNN models best suited for mobile and edge devices for on-device medical image analysis which can decrease the dependency over the cloud-based processing. Through federated learning, improvement in the model is ensured without loss of privacy of patient data, and through a multi-modal diagnostic approach, which uses image, audio, and analysis based on symptoms, the accuracy is improved. The system is also based on efficient image compression and quantization techniques to enable real-time diagnosis in bandwidth-constrained environments. Also, for the safe and tamper-proof management of medical records, blockchain technology is added to the healthcare networks that can be decentralized regarding data integrity and compatibility with one another. It has been designed to work in off-grid environments powered by AI stations, solar powered enabling it to be sustainable to rural or underdeveloped regions. Diagnosis is also shown to be highly accurate, have low latency, and be adaptable for different medical conditions in experimental evaluation. It is this type of scalable, cost-effective, and privacy-preserving solution to improve healthcare accessibility and diagnostic burden in resource-limited settings that this research provides. Future work involves expanding the system for more general disease detection and deployment in a real environment.
A Comparison of Radial Basis Function and Multilayer Perceptron Network as tool for classification of Medical Data S. Leoni Sharmila, C. Dharuman, P. Venkatesan Journal of Physics Conference Series, 2019 Artificial Neural Network has become a popular tool in developing systems that encircles human proficiency. The importance of exact detection is exceptionally important for proper treatment and preserve of disease. Clinical cytology has improved tremendously in disease diagnosis. In this paper, two Artificial Neural Network (ANN) methods, Multilayer Perceptron (MLP) and Radial Basis Function (RBF) network are compared. The RBF network predicts comparatively better accuracy in compared to MLP methods. Also it was detected that the RBF method requires a lesser amount of time for the development of the model, this is because there is no repetition to reach the favourable parameters in the model.
A fuzzy based classification – An experimental analysis S. Leoni Sharmila, C. Dharuman, P. Venkatesan International Journal of Innovative Technology and Exploring Engineering, 2019 Soft Computing has become popular in developing systems that encloses human expertise. Imaging technologies and clinical cytology has improved in disease diagnosis. Exact detection is extremely important for proper treatment and cure of disease. Two soft computing technique Neural Network and Support Vector Machine are used for classification of Caridotocography data set. This paper clearly explains the advantages of hybrid technique, when Fuzzy is combined with Neural Network and Support Vector Machine it is clearly noticed that there is an increase in accuracy of classification rate.
A novel neuro-fuzzy system for classification Global Journal of Pure and Applied Mathematics, 2016
