Multi-medical image protection: compression–encryption scheme based on TLNN and mask cubes

With the massive use of online medical image data, the protection of the private information of patients has become an important issue to be addressed. A tabu learning neuron network (TLNN) and mask cubes produced from the original images based medical multi-image compression–encryption (MICE) scheme are designed to solve this issue. Firstly, multiple 2D images are compressed by using compression sensing (CS) technology, then the compressed images are integrated into eight small cubes, considering them as 3D cubes. Next, chaotic sequences iterated by the TLNN are used to integrate the small cubes into one large cube, which is subjected to confusion operations by using the 3D shuffling algorithm. And then, two complementary mask cubes are generated with the large cube, and they are used to perform different diffusion operations on the large cube to obtain two new image cubes. Finally, the two image cubes are combined to complete the MICE scheme. The performance test results nicely prove the compatibility, efficiency and safety of the scheme, and guarantee the security of medical images.