English

Reconstructing random pictures

Combinatorics 2025-01-29 v3 Probability

Abstract

Given a random binary picture PnP_n of size nn, i.e., an n×nn\times n grid filled with zeros and ones uniformly at random, when is it possible to reconstruct PnP_n from its kk-deck, i.e., the multiset of all its k×kk\times k subgrids? We demonstrate ``two-point concentration'' for the reconstruction threshold by showing that there is an integer kc(n)(2logn)1/2k_c(n) \sim (2 \log n)^{1/2} such that if k>kck > k_c, then PnP_n is reconstructible from its kk-deck with high probability, and if k<kck < k_c, then with high probability, it is impossible to reconstruct PnP_n from its kk-deck. The proof of this result uses a combination of interface-exploration arguments and entropic arguments.

Keywords

Cite

@article{arxiv.2210.09410,
  title  = {Reconstructing random pictures},
  author = {Bhargav Narayanan and Corrine Yap},
  journal= {arXiv preprint arXiv:2210.09410},
  year   = {2025}
}

Comments

v3: 9 figures, 23 pages, substantial additions made to the presentation of main proofs; final version appearing in RSA

R2 v1 2026-06-28T03:51:44.133Z