Answers are grounded in this paper and cite their sources.
This analysis demonstrates enhanced uniform distribution in random sequences from webcams, suggesting improvements over software-based methods.
At the current stage of development, software-based random number generation is increasingly vulnerable to attacks due to the growing computational power of modern systems. Hardware-based generation relies on reliable stochastic physical phenomena, yet often demonstrates low throughput or an insufficient level of statistical quality. In this study, linear cellular automata are utilized as a post-processing mechanism aimed at improving key statistical properties of the random number sequences. The study presents the results of improving the statistical properties of number sequences obtained from a standard webcam with regard to one of the key cryptographic requirements – uniform distribution of values. Prior analyses revealed that inherent stochastic behavior within webcam sensors produces irregular, non-uniform distributions in the output sequences. This limitation can be mitigated by post-processing using linear cellular automata, specifically employing rules 30, 90, 105, and 150. To evaluate the effectiveness of this method, a comparative analysis was conducted against software-generated sequences produced via the Java SecureRandom class. It is shown that by applying a sufficient number of iterations, it is possible to achieve the required level of uniformity in the distribution of sequence elements. However, the generated sequences did not fully pass the NIST test suite. Statistical uniformity was assessed efficiently using lightweight Java-based libraries, enabling rapid integration even on low-performance hardware such as smartphones. The research findings can be applied in the design of a hardware-based random number generator.
No takes yet. Share an insight, caveat, or question.
Prokhorov et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: