Research On Multimodal Digital Human Identity Security Protection Technology Based On Blockchain

With the flourishing development of new technologies such as AI,virtual reality and deep learning,virtual digital humans have been greatly improved in all aspects of performance and more efficient in the production process,and have evolved from the digitalisation of the appearance of virtual digital humans to the development of intelligent human-computer interaction.In many business scenarios,virtual anchors and virtual employees can be seen in our lives,and play an important role in the fields of cultural tourism,film and television,finance and media.Multimodal digital humans are the latest research in the field of AI and are based on natural language processing,speech and vision technologies.The multimodal digital human can be functionally divided into content broadcasting type and dialogue interaction type digital human.In terms of image,multimodal digital human can be divided into 2D digital human and 3D digital human.2D digital human is a real person-driven virtual digital human,while 3D digital human is similar to cartoon digital human,and thesi paper mainly focuses on 2D digital human for identity security research.At present,while multimodal digital human technology is rapidly iterating and commercializing,major AI digital human vendors do not have a good method to protect the identity security of multimodal digital humans.When AI digital human vendors' customers generate broadcast-type services through multimodal digital humans,such as generating news broadcasts and anti-fraud promotional videos through text,and uploading the videos to internet platforms.However,the dissemination and regulation of information on internet platforms is quite a difficult thing to do,while the development of AI technology allows for tampering with broadcast content through AI face-changing technology,thus causing multimodal book digital people to face problems such as copyright disputes and difficulties in maintaining rights.It also causes economic losses and affects the development of multimodal digital human technology to AI digital human manufacturers.The main work of thesis paper,which addresses the study of multimodal digital human identity security,is as follows:(1)In response to the problems of high communication complexity and scalability of the PBFT consensus algorithm in coalition chains,thesis paper proposes a low-cost consensus algorithm that combines the consensus efficiency and fault tolerance features of Raft and PBFT consensus algorithms,while using a BLS signature scheme to optimize the consensus process.(2)thesis paper designs an identity security protection scheme based on blockchain technology,which can solve the identity security problems encountered by multimodal digital people in commercial applications by applying the traceability and tamper-proof features of blockchain technology.The scheme is implemented based on the underlying technical architecture of the federated chain Hyperledger Fabric,while the proposed low-cost consensus algorithm is applied to the Hyperledger Fabric technical architecture to improve consensus efficiency.(3)A network model for data storage under the blockchain chain is designed,with multimodal digital human audio and video data stored on the IPFS distributed storage file system.with content addressing and tamper-proof features,IPFS can effectively solve the problems of resource wastage and performance degradation caused by storage on the blockchain chain.Innovations in thesis paper:(1)A low cost consensus algorithm is proposed,which optimizes the consensus process using the BLS signature algorithm,reduces the communication complexity of the PBFT algorithm and the RBFT algorithm,and reduces the network communication volume.In the same network environment,through comparison experiments with PBFT and RBFT consensus algorithms,this consensus algorithm performs better in terms of consensus latency,communication overhead and throughput,and can effectively break the performance bottleneck of current coalition chains using PBFT consensus algorithm.(2)A blockchain-based digital human identity security protection scheme is proposed,which applies blockchain technology,digital signature technology and encryption technology to ensure that multimodal digital person identity information cannot be tampered with,and also has the advantages of open and transparent information,traceability and trustworthiness,as well as data storage security,thus ensuring that multimodal digital person identity information is secure and reliable.