Search by tag «Journal of Molecular Liquids» 1 results

• Classification From ITMO To Help AI Generate Random Noise on Request

  Random noise is useful for generating truly random numbers, such as for stronger passwords and data protection, but it’s also what helps reconstruct full-scale images obtained from small cameras placed within the body during medical screenings. Using machine learning and fractal analysis methods, researchers from ITMO University have studied a source of random noise – the textures inside a layer of liquid crystal – and created a classification based on the analysis of large- and small-scale fractal dimensions. Using this data, it's possible to train a neural network to select the parameters for random noise generation, adjusting it for various purposes. The results of the study have been published in Journal of Molecular Liquids.

  21.11.2024