Kubernetes-based Deep Learning with Blockchain for Cancer Image Prediction

Abstract

The ability of modern artificial intelligence techniques to analyze complex data, including medical images, with greater precision and depth has significantly increased the amount of real-world data that can be collected for scientific research. This has led to the development of a global library and elevated computing choices to meet the demand for training model calculation. The development of medical image biomarkers involves additional processes that don't necessitate spending so much money on expensive computer resources. To collect and manage cancer data, this study presents a Blockchain-enabled deep learning framework for the Kubernetes cluster which replicates the training set of parameters of DL techniques across remote cluster nodes. Healthcare providers' data will be synchronized thanks to the distributed storage system known as the blockchain. Data are gathered for this suggested framework utilizing the Kaggle Dataset from several hospitals. Data preprocessing shows pixel correction and data augmentation. Following data preprocessing, U2-Net segmentation is used to segment up the images. This article uses blockchain to decentralize data, security, and privacy. A system called Kubernetes cluster is used to manage diverse containerized applications, Workload Distribution, Scalability, and Storage Management. The Kubernetes cluster is revealed in the study when investigating the cancer prediction capabilities of the hybrid learning algorithms VGG19 and BiLSTM. This proposed paper forecasts the article and explains the various applications for breast, kidney, and oral cancer predicated on the Blockchain-Enabled deep Learning model on Kubernetes and providing authentication for the use of the majority of the platform's features using a provided case.