Low Resolution MRI Images Privacy Feature Mapping and Classification Using LSTM-CNN Models in IoMT Healthcare Applications

• Published in: SN computer science Vol. 5; no. 7; p. 880
• Main Authors: R, Sindhuja, Kapse, Arvind S.
• Format: Journal Article
• Language: English
• Published: Singapore Springer Nature Singapore 14.09.2024; Springer Nature B.V
• Subjects: Advances in Computational Approaches for Image Processing; Artificial intelligence; Artificial neural networks; Bandwidths; Classification; Cloud Applications and Network Security; Communication channels; Computation; Computer Imaging; Computer Science; Computer Systems Organization and Communication Networks; Data Structures and Information Theory; Data transmission; Datasets; Decision making; Image resolution; Information Systems and Communication Service; Internet of medical things; Literature reviews; Magnetic resonance imaging; Mathematical models; Medical equipment; Medical imaging; Neural networks; Optimization techniques; Original Research; Pattern Recognition and Graphics; Privacy; Quality standards; Software Engineering/Programming and Operating Systems; Standardization; Third party; Vision; Wireless Networks; CNN; LSTM; MRI computing; Medical IoT; Medical data processing; IoT
• ISSN: 2661-8907; 2662-995X; 2661-8907
• DOI: 10.1007/s42979-024-03230-4

Medical images classification and decision making via Internet of Medical Things (IoMT) applications is a challenging task. The qualities of medical images are optimized and lower resolution data is transferred via the communication channel to IoT servers, making the decision support a complicated computing process. In this paper, a novel convolutional neural networking (CNN) model and long short term memory (LSTM) model based decision making of low resolution medical images is computed. The framework is generalized for MRI datasets, the CNN + LSTM combination extracts features from low resolution medical images and further classifies the order of MRI application based on the thresholding feature of host application. The technique is cross-validated with multiple MRI dataset samples for performance estimation. The interim decision making and integrated learning model provides the framework an added efficiency on optimized computing.