Non-fungible tokens (NFTs) have emerged as a unique form of digital asset, ranging from digital art to real-world commodities. While NFTs offer security through decentralization and smart contract enforcement, they remain vulnerable to cyber threats. This paper proposes an enhanced cryptographic framework designed to secure the transfer of NFTs, with a specific focus on image-based NFTs. The system integrates a novel encryption algorithm that uses pixel-based encryption techniques alongside a Generative Adversarial Network (GAN) to generate random images, which are further encrypted with Gaussian noise to enhance their resilience against unauthorized decryption. The proposed system ensures secure image transfers by utilizing off-chain methods for efficient and safe data transmission. The decryption process requires the recipient to possess a private identification, transaction key, and off-chain data to ensure exclusive access. Experimental results comparing various encryption algorithms demonstrate that the proposed method provides strong encryption with minimal execution time, ensuring a secure transfer process and protecting digital assets in the evolving digital ecosystem.

Keywords: Non-Fungible Tokens (NFTs), Generative Adversarial Networks (GANs), Image Encryption, Cryptography, Pixel-based Encryption, Gaussian Noise, Decryption Protocol.