Current browse context:
cs.IT
Change to browse by:
References & Citations
Computer Science > Information Theory
Title: Estimation in Poisson Noise: Properties of the Conditional Mean Estimator
(Submitted on 9 Nov 2019 (v1), last revised 31 Mar 2020 (this version, v2))
Abstract: This paper considers estimation of a random variable in Poisson noise with signal scaling coefficient and dark current as explicit parameters of the noise model. Specifically, the paper focuses on properties of the conditional mean estimator as a function of the scaling coefficient, the dark current parameter, the distribution of the input random variable and channel realizations. With respect to the scaling coefficient and the dark current, several identities in terms of derivatives are established. For example, it is shown that the gradient of the conditional mean estimator with respect to the scaling coefficient and dark current parameter is proportional to the conditional variance. Moreover, a score function is proposed and a Tweedie-like formula for the conditional expectation is recovered. With respect to the distribution, several regularity conditions are shown. For instance, it is shown that the conditional mean estimator uniquely determines the input distribution. Moreover, it is shown that if the conditional expectation is close to a linear function in terms of mean squared error, then the input distribution is approximately gamma in the L\'evy distance. Furthermore, sufficient and necessary conditions for linearity are found. Interestingly, it is shown that the conditional mean estimator cannot be linear when the dark current parameter of the Poisson noise is non-zero.
Submission history
From: Alex Dytso [view email][v1] Sat, 9 Nov 2019 17:59:10 GMT (70kb)
[v2] Tue, 31 Mar 2020 15:06:56 GMT (275kb)
Link back to: arXiv, form interface, contact.