Post-Processing. A round of experiment is done when all parties send the fully passive signals and re- ceive Xxxxxxx’s detection announcement. After a large number of rounds are performed, the first party, A0, pub- licly announce a random binary sequence that decides if the corresponding round is a key generation (KG) round or a parameter estimation (PE) round, with probabili- ties pX and 1 pX , respectively. One key advantage of our fully passive CKA protocol is that there is no basis sifting across multiple users, as each other user simply chooses the same publicly-announced basis as Xxxxx after the transmission. This is because, for our protocol, both the KG and PE round use the same physical signal, and the difference lies only in post-processing, which can be done after the transmission. This is especially advanta- geous when the number of users is large. For NU users, the advantage of sifting efficiency over an active CKA scheme is pNU −1. 2a) KG round: Each party, Ai, divides the whole 2π range into M slices where M must be an even number, and each of their local phase measurements should land into one of the slices indexed ki1 and ki2. A KG round is successful only when the slice indices obey ki1 = ki2 = 1 for all i, or ki1 = ki2 = M/2 + 1 for all i. The top slice corresponds to the plus state in X-basis, and hence corresponds to a classical bit 0, and the bottom slice cor- responds to the minus state in X-basis, and hence corre- sponds to a classical bit 1. The unsuccessful events are discarded. This post-selection method discards the ma- jority of the signals, resulting in very low sifting. How- ever, we will show in the following sections that the dis- carded signals can contribute to the key rate. For now, to introduce the protocol clearly, we do not yet consider other signals. 2b) PE round: Each party, Ai, calculates their local phase differences, ϕi1 ϕi2 , this is equivalent to find- ing their output signal intensity in the range between 0 and umax. For example, in a 3-decoy protocol, each party divides their intensity range into 3 parts. Each party announces which part their signal falls into based on ϕi1 ϕi2 . In this way, all n parties essentially di- vide an n-dimensional square into small squares, each square, labeled Sijkl, corresponds to a decoy state choice by all parties. They perform decoy-state analysis to es- timate the yield, Yn1n2n3n4 , and subsequently calculate
Post-Processing. (1) All delivered images will have at least the base post-processing applied.
(2) Basic post-processing is considered:
a) Brightness correction b) Contrast correction c) Color or black-and-white correction d) White balance correction
Post-Processing. The CONSULTANT shall be responsible for all post- processing of all data.
Post-Processing
