Security and privacy considerations in blockchain-based computer vision

Abstract

This chapter reviews the security and privacy issues related to the integration of blockchain and computer vision (CV) systems. CV is applied in the health sector, in car driving, in shopping centers, and in surveillance, but it highly depends on image and video data, which are difficult to authenticate, secure, or even kept private. The decentralized, immutable, and transparent format of blockchain provides a good choice to address these issues as it allows sharing of data safely, controlling access to data, and ensuring its auditability. This chapter starts with the introduction to the concept of CV applications and the confidentiality risks of these applications and then proceeds to explain how the concept of blockchain can be used to create a dependable, tamper-proof system to handle visual data, through the use of smart contracts and consensus algorithms. It talks of state-of-the-art methods, such as encryption, zero-knowledge, and federated learning (FL), to guarantee privacy preservation in blockchain-based CV systems. In fact, the real-life application illustrates how blockchain is used to lock up medical imaging, self-driving vehicles, and surveillance data. Lastly, this chapter discusses new regulatory and ethical issues such as ownership of the data, legislative privacy, and ethical smart use of surveillance technology.