Imaging Through Noise With Quantum Illumination Data Analysis Readme All frames saved in the archive are 128x128 pixels. UoG data analysis: Data in archive named Data_UoG.7z Thermal data used in the generation of Figures: S3-S4, S6-S7 found in folder: QI0710ThermalData Losses data used in the generation of Figures: 5 and S5 found in folder: QI0710LossesData 1. For each frame take the two beams as regions of interest (ROIs) such that the correlation peak is centered on the central pixel 2. Rotate one by an angle of pi (or equivalently invert in x and y) 3. Perform an AND operation and use result to build the AND-images 4. Assess contrast using the Michelson Contrast with mask: Masks_UoG/UoGMask0710Data.txt to define the UoG or bright region and mask: Masks_UoG/BGMask0710Data.txt to define the background or dark region. 5. Take ratio of contrasts Vq / Vc to evaluate the quantum illumination advantage A Assess contrast on images constructed from 100,000 frames to get mean and standard error on the mean Bird in a Cage, Fish in a Net analysis used in the generation of Figures 3 and 4: Data in archive named StructuredBackgroundData.7z 1. For each frame take the two beams as regions of interest (ROIs) such that the correlation peak is centered on the central pixel 2. Rotate one by an angle of pi (or equivalently invert in x and y) 3. Perform an AND operation and use result to build the AND-images 4. Mask the image using StructuredBackgroundData\Masks\BackgroundMask.txt and calculate the mean value of the regions of the image that do not comprise either the bird or the cage. This value corresponds to the floor and the other values below are relative to this so as to assess the relative brightnesses of the bird and the cage to calculate the noise rejection ratio and the distinguishability metric. 5. Using StructuredBackgroundData\Masks\ClassicalCageMask.txt StructuredBackgroundData\Masks\QuantumBirdMask.txt, calculate the mean value of the classical cage , and the mean value and standard deviation of the quantum bird regions and sigma_O. 6. Assess noise rejection ratio (NRR) by taking of regions containing only the quantum illuminated object and regions containing only the classically illuminated cage . Compare the ratio for the classical and quantum illumination AND-image according to the text and equation 1 in the manuscript. 7. Assess distinguishability ratio (D) by taking the ratio of regions containing only the quantum illuminated object and regions containing only the classically illuminated cage plus the standard deviation on the regions of the object sigma_O. Compare the ratio for the classical and quantum illumination AND-image according to the text and equation 2 in the manuscript. Bit Error Rate (BER) analysis: Data found in folder: QI0710ThermalData Blind Strategy: 1. Calculate the BER by summing the UoG ROIs for a number of frames (creating the classical image) and using masks: Masks_UoG/UoGMask0710Data.txt and Masks_UoG/BGMask0710Data.txt for the bright and dark regions respectively to determine the number of background pixels that are bright and the number of UoG or object pixels that are dark. 2. Find the number of frames N (to the nearest block of 50) that minimises the bit error rate for the classical data. 3. Assess the BER for the quantum illumination AND-image calculated on N / (p+d) frames so that the same number of events feature in the images on which the BER is to be assessed. 4. calculate the BER for each of the data sets of different thermal illumination levels. Known noise: 1. Weighted sum of the quantum illumination AND-image and the classical image 2. mimimise BER by changing the value of the weighting factor Peak Data: Used to generate figure 2. A cross-correlation is performed between the two beams and the peak is fitted by a gausian in the x and y directions. The stated value of the peak width is the mean of the standard deviation of the peak in these two directions.