In recent days, the Internet of Medical Things (IoMT) is commonly employed in different aspects of healthcare applications. Owing to the increasing necessitates of IoT, a huge amount of sensing data is collected from distinct IoT gadgets. To investigate the generated data, artificial intelligence (AI) models plays an important role to achieve scalability and accurate examination in real-time environment. However, the characteristics of IoMT result in certain design challenges, namely, security and privacy, resource limitation, and inadequate training data. At the same time, blockchain, an upcoming technology, has offered a decentralized architecture, which gives secured data transmission and resources to distinct nodes of the IoT environment and is stimulated for eliminating centralized management and eliminates the challenges involved in it. This paper designs deep learning (DL) with blockchain-assisted secure image transmission and diagnosis model for the IoMT environment. The presented model comprises a few processes namely data collection, secure transaction, hash value encryption, and data classification. Primarily, elliptic curve cryptography (ECC) is applied, and the optimal key generation of ECC takes place using hybridization of grasshopper with fruit fly optimization (GO-FFO) algorithm. Then, the neighborhood indexing sequence (NIS) with burrow wheeler transform (BWT), called NIS-BWT, is employed to encrypt the hash values. At last, a deep belief network (DBN) is utilized for the classification process to diagnose the existence of disease. An extensive experimental validation takes place to determine the analysis of the optimal results of the presented model, and the results are investigated under diverse aspects.

近年来，医疗物联网（IoMT）通常用于医疗保健应用程序的不同方面。由于物联网的需求不断增长，因此从不同的物联网小工具中收集了大量的传感数据。为了调查生成的数据，人工智能（AI）模型在实时环境中实现可伸缩性和精确检查方面起着重要作用。但是，IoMT的特性导致某些设计挑战，即安全性和隐私性，资源限制以及训练数据不足。同时，即将到来的技术区块链提供了一种去中心化的架构，该架构可将安全的数据传输和资源提供给IoT环境的不同节点，并受到激励，消除了集中管理并消除了其中涉及的挑战。本文针对IoMT环境设计了基于区块链的安全图像传输和诊断模型的深度学习（DL）。提出的模型包括几个过程，即数据收集，安全交易，哈希值加密和数据分类。首先，应用椭圆曲线密码学（ECC），并通过将蚱hopper与果蝇优化算法（GO-FFO）进行杂交来实现ECC的最佳密钥生成。然后，采用具有挖土车变换（BWT）的邻域索引序列（NIS），称为NIS-BWT，对哈希值进行加密。最后，将深度信念网络（DBN）用于分类过程以诊断疾病的存在。进行了广泛的实验验证，以确定对所提供模型的最佳结果的分析，