Ultra‐Wide‐Field Noninvasive Imaging Through Scattering Media Via Physics‐Guided Deep Learning

Lintao Peng, Mingwei He, Jeff Zhu, Sujit K Sahoo, Liheng Bian, Cuong Dang · 2026

Read Ultra‐Wide‐Field Noninvasive Imaging Through Scattering Media Via Physics‐Guided Deep Learning by Lintao Peng, Mingwei He, Jeff Zhu, Sujit K Sahoo, Liheng Bian, Cuong Dang (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13335759.txt.

Description

Lintao Peng and co-authors present, in a 2026 study, a method to image through frosted, scattering stuff across a very wide field. Light that hits such stuff spreads as speckle, and past methods could read just a thin slice of the scene. Their model, UNI-Net, is guided by physics at each step. A physics model first makes fake speckle data at scale, which cuts the need for real lab data by a factor of ten. Each speckle frame is then split into patch channels that guide a diffusion build.

A new block reads space and channel traits side by side at low cost. Scores hit 31.23 decibels at 41 times the base view range, about half more than past best. At 164 times that range, where past methods fail, it still rebuilt scenes at 27.21. The work cuts both data needs and compute cost.

Catalog record

CREATORLintao Peng, Mingwei He, Jeff Zhu, Sujit K Sahoo, Liheng Bian, Cuong Dang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE54 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC13335759

Browse the full library · About the Tower of Babel