{"id":"3d6781d7-2562-43cf-a0dc-1a67da1681b2","arxiv_id":"2504.12092","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Line-scanning 'virtually structured detection' is applied to a Rydberg-atom terahertz imager, improving resolution by 74(3)% at 0.55 THz without deconvolution.","lead":"This paper demonstrates structured illumination super-resolution imaging at 0.55 THz using a line-scanning approach called virtually structured detection with an atomic-vapor terahertz camera. The method sharpens images of transmission targets by about 74 percent without deconvolution, extending a known optical microscopy technique into the terahertz band.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The quantitative 74(3)% claim rests on a Gaussian-erf fit to one edge in a non-deconvolved reconstruction; ringing or neighboring-bar leakage could bias the fitted ESF width.","rationale":"The paper demonstrates a genuine qualitative effect: vertical bars at 1.12 lp/mm that are unresolved in the widefield image appear resolved in the VSD image, and the Fourier spectra show extended support along the scan direction. So the qualitative claim of super-resolution is plausible and independently supported by the images. The load-bearing weakness is exactly where the quantitative claim is generated: a single-edge Gaussian-erf fit on a reconstruction whose OTF is not flat and likely has sidelobes. This can bias the inferred FWHM, and the paper provides no error bar for edge-to-edge variability or model mismatch. The reader's verdict of CONDITIONAL captures this correctly: the result is not disproven, but the headline 74(3)% should not be accepted as a robust quantitative resolution measure without a direct, multi-edge, model-independent check. My stress-test agrees with the reader's weakest assumption and does not move the verdict.","tokens_in":8249,"tokens_out":8490,"duration_ms":95437,"concrete_test":"From the published Fig. 3E profiles (or the raw scan stack, if made available), measure the 10-90% edge-rise width on every resolvable edge in Element 2 without any error-function fit, and compare the widefield-to-super-resolution ratio with the claimed widefield FWHM / SR FWHM = 0.94/0.54 = 1.74. In parallel, run the paper's Gaussian-erf fitting pipeline on synthetic edges formed by convolving an ideal step with the expected OTF (including ringing) to calibrate the bias of that fitting procedure. If the direct edge-width ratio or the synthetic calibration indicates a deviation larger than 10% from 74%, the headline resolution number should be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central figure of merit is the PSF FWHM inferred from the edge profile in Fig. 3E (Section 4, paragraph beginning 'To quantify the improvement...'). The super-resolution image is formed by summing shifted Fourier bands without deconvolution (Eqs. 4-5), so the effective OTF is a sum of shifted widefield OTFs rather than a smooth, positive-definite kernel. Such an OTF generically produces edge overshoot and ringing. Fitting a Gaussian error function to an edge with ringing can return a narrower 'PSF' than the true resolution, because the fit absorbs sidelobes into a steeper central slope. The profile is also extracted from a periodic USAF bar pattern, so the assumed isolated-step ESF is not exact: the neighboring bar contributes an opposite edge within the fit window. The quoted uncertainties (+/-0.01 mm) are fit-parameter errors, not edge-to-edge or model-choice variation. Moreover, the paper does not report the modulation frequency p, the ROI width, or the scan step used in the reconstruction, so the effective OTF and the edge window cannot be independently checked. Consequently the 74% improvement is not yet established as a system resolution gain.","agreement_with_reader":"agree"},"referee_report":null,"author_rebuttal":null,"desk_editor":null,"rs_alignment":null,"lean_confirmation":null,"pith_extraction":null,"created_at":"2026-08-16T12:38:36.762889+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":null,"supporting_citations":[],"review_version":1}