Homomorphic-Encrypted Volume Rendering

Computationally demanding tasks are typically calculated in dedicated data centers, and real-time visualizations also follow this trend. Some rendering tasks, however, require the highest level of confidentiality so that no other party, besides owner, can read or see sensitive data. Here we present a direct volume approach performs directly on encrypted by using homomorphic Paillier encryption algorithm. This ensures rendered image uninterpretable to server. Our pipeline introduces novel approaches for encrypted-data compositing, interpolation, opacity modulation, as well simple transfer function design, where each these routines maintains privacy. We performance memory overhead analysis is associated with our privacy-preserving scheme. open secure opposed through obscurity. Owners only have keep their key confidential guarantee privacy images. work is, knowledge, first remote volume-rendering does not any server involved be trustworthy; even cases when compromised, will leaked foreign party.