.png){kind=icon}
.png){kind=icon}
Return
Optical cryptography based on computational ghost imaging and computer-generated holography
.png){kind=icon}
PRE
AI
DOI：10.1016/j.optlaseng.2024.108780.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
We present a method for jointly utilizing computational ghost imaging (CGI) and computer-generated holography (CGH) in optical encryption via deep learning. Two distinct secret images are independently encrypted by CGI and CGH, and subsequently combined into a single ciphertext, where the optically collected bucket signals need to be binarized and sparsified. The ciphertext is embedded into the host image using a deep learning network. In this approach, the networks used to embed the ciphertext into the host image and to extract the ciphertext from the stego image are trained collaboratively. This work addresses a drawback of traditional CGIbased encryption, which requires a large number of illumination patterns during transmission and decryption. Furthermore, this is an attempt to decrypt and restore plaintext from sparse and binarized bucket signals, and it is also the application of non-direct and direct imaging for dual-image encryption.
Keywords:
Optical encryption
Ghost imaging
Computer-generated holography
Deep learning
Journal
IF:
3.7
Papers:
7.2K
Citations:
1.7W