| In recent years,with the rapid development of diversified social platforms,users have become increasingly enthusiastic about sharing what they see and hear about their lives through the internet.But currently,on these social platforms,the phenomenon of copyright infringement often occurs,making it difficult to verify the rights of works.Moreover,based on the current implemented copyright protection systems,there is no copyright protection for digital works that integrate text and images or text and videos as a whole.This article starts from the problems existing in current social platforms,and combines the characteristics of blockchain and deep learning technology to design and implement a blockchain based digital work copyright protection system.The digital work includes text plus images and text plus videos.Firstly,a deep learning algorithm is used to extract the feature vectors of works for copyright acquisition,in order to obtain robust copyright features.In this thesis,an improved Swin Transformer model algorithm is used to extract features from images and video keyframes in digital works,and BERT model is used to extract features from text,The extracted feature vectors are stored as copyright information of the work in the Milvus database for protection,and combined with Milvus’ similarity retrieval function,the problem of copyright infringement of the work is solved.Secondly,this article conducts a requirement analysis of a blockchain based digital work copyright protection system by conducting research on current social network platforms and conducting in-depth analysis of blockchain technology.The overall architecture and functional modules of the system are designed,and a detailed process design is carried out for the login module,work copyright management module,copyright application processing module,copyright appeal processing module,copyright authorization processing module,and other modules.Finally,this paper adopts a hierarchical architecture model for system development,uses Vue.js+Spring Boot+FastAPI framework to complete the development of a digital works copyright protection system based on the Ethereum blockchain,and uses Solidity language to complete the smart contract for works copyright uplink and query functions;The front-end developed through Vue.js embedding Element-UI provides RESTful API interface communication with the Spring Boot backend service;The Spring Boot service interacts with Ethereum on chain contracts through Web3j,enabling on chain and off chain interactions;Provide functional interfaces such as vector storage and similarity retrieval to the backend of the system through the FastAPI framework,deep learning model,and Milvus database interaction.Finally,it underwent system testing to ensure the safety,correctness,and scalability of the system’s functions. |