Abstract
Assessing that a probability of false alarm is below a given significance level is a crucial issue in watermarking. We pro- pose an iterative and self-adapting algorithm which estimates very low probabilities of error. Some experimental investiga- tions validates its performance for a rare detection scenario where there exists a close form formula of the probability of false alarm. Our algorithm appears to be much quicker and more accurate than a classical Monte Carlo estimator. It even allows the experimental measurement of error exponents.