A secured biomedical image processing scheme to detect pneumonia disease using dynamic learning principles

Abstract

Now-a-days, the medical industry is growing a lot with the adaptation of latest technologies as well as the logical evaluation and security norms provides a robust platform to enhance the effectiveness of the industry at a drastic level. In this paper, a digital bio-medical image processing based Pneumonia disease identification system is introduced with enhanced security features. Due to improving the efficiency of the application, a well-known watermarking based security constraint is included to provide the protection to the respective hospital environment and patients as well. To avoid these issues, some sort of security aspects need to be followed so that this paper included watermarking based security to provide a rich level of protection to the images going to be tested. The main intention of this paper is to introduce a novel security enabled digital image processing scheme to identify the Pneumonic disease in earlier stages with respect to the proper classification principles. In this paper, a novel deep learning algorithm is introduced called enhanced Dynamic Learning Neural Network in which it is a hybrid algorithm with the combinations of conventional DLNN algorithm and the Support Vector Classification algorithm. This proposed approach effectively identifies the Pneumonia disease in earlier stages but the security inspection on the testing stage is so important to analyze the disease. The respective testing image is properly watermarked with the logo of the corresponding hospital; the image is processed otherwise the proposed approach skips the image to process. These kinds of security features emphasize the medical industry and boost up the levels more as well as the patients can get an appropriate error free care with the help of such technology. A proper Chest X-Ray based Kaggle dataset is considered to process the system as well as which contains 5856 Chest X-Ray images under two different categories such as Pneumonia and Normal. With respect to processing these images and identifying the Pneumonia disease effectively as well as the proposed watermarking enabled security features provide a good impact in the medical field protection system. The resulting section provides the proper proof to the effectiveness of the proposed approach and its prediction efficiency.