Lossy compression of weak lensing data

TL;DR

This paper evaluates a lossy compression algorithm for space-based weak lensing data, demonstrating it reduces data size significantly with minimal bias and noise increase, thus aiding large sky surveys for dark matter and dark energy studies.

Contribution

It provides a detailed assessment of a specific lossy compression method's impact on weak lensing measurements, showing minimal bias and noise effects.

Findings

Achieves 6.7x compression of images with minimal bias

Introduces negligible bias in ellipticity measurements after correction

Demonstrates no bias in sky background after compression

Abstract

Future orbiting observatories will survey large areas of sky in order to constrain the physics of dark matter and dark energy using weak gravitational lensing and other methods. Lossy compression of the resultant data will improve the cost and feasibility of transmitting the images through the space communication network. We evaluate the consequences of the lossy compression algorithm of Bernstein et al. (2010) for the high-precision measurement of weak-lensing galaxy ellipticities. This square-root algorithm compresses each pixel independently, and the information discarded is by construction less than the Poisson error from photon shot noise. For simulated space-based images (without cosmic rays) digitized to the typical 16 bits per pixel, application of the lossy compression followed by image-wise lossless compression yields images with only 2.4 bits per pixel, a factor of 6.7 compression. We demonstrate that this compression introduces no bias in the sky background. The compression introduces a small amount of additional digitization noise to the images, and we demonstrate a corresponding small increase in ellipticity measurement noise. The ellipticity measurement method is biased by the addition of noise, so the additional digitization noise is expected to induce a multiplicative bias on the galaxies' measured ellipticities. After correcting for this known noise-induced bias, we find a residual multiplicative ellipticity bias of m ~ -4x10^{-4}. This bias is small when compared to the many other issues that precision weak lensing surveys must confront, and furthermore we expect it to be reduced further with better calibration of ellipticity measurement methods.