Second-Order Image Characterization with Distortion and Secrecy

Abstract

Image-size characterization connects coding constraints to the probability carried by source cells and their output images. At second order, this connection must retain the joint fluctuations of all active constraints and distinguish encoder-observable source variation from unobserved channel noise. We extend the single-output posterior characterization to simultaneous posterior and conditional-composition image bounds. The bounds preserve the actual coupling of multiple outputs, quantify conditional likelihood losses for irregular cells, and admit actual-distortion marks. Two further extensions replace disjoint counting by weighted reconstruction-policy images and replace output cardinality by posterior-mass constraints for secrecy. Direct applications yield matching second-order regions for compatible multi-source helper intersections, one lossy reconstruction with several informed lossless reconstructions, Gaussian common/private descriptions, and privacy amplification from a prescribed conditional type. The Gaussian description result also gives joint rate--distortion offsets. For general helper sections, genuinely lossy informed reconstruction, Gaussian Wyner--Ziv coding, and reconciliation followed by extraction, the same image methods give explicit rate bounds with their remaining losses identified. The resulting rate corrections reflect both the image constraints and the observations available to each terminal.