Internet of Medical Things (IoMT) is a study field that is in high demand and is used in most medical applications. When dealing with medical data or images on a decentralized site, security is a hard problem to solve. To improve the security of medical images in IoMT, a deep learning-based blockchain framework that works well and has lower transaction costs is suggested. The suggested study has four steps: getting the image, encrypting it, finding the best key, and storing it safely. In the picture acquisition stage, the input images are gathered in the first step. Then, the gathered medical images are encrypted using coupled map lattice (CML). This encryption process helps keep medical pictures from being seen by attackers. The optimal keys are made by using the opposition-based sparrow search optimization (O-SSO) method. This makes the encrypted images more private. Distributed ledger technology (DLT) and smart contract-based blockchain technology are used to store these pictures that have been encrypted. This blockchain technology improves the integrity and authenticity of the data and makes it possible to send medical pictures securely. In the classification step, the disease is found by using the proposed Recurrent Generative Neural Network (RGNN) model. This is done after the image has been decrypted. For the simulation analysis in the suggested study, a Python tool was used, and the medical images came from CT images in the COVID-19 dataset