Abstract
The SynthOCT Challenge 2026 asks participants to construct an optical coherence tomography (OCT) digital phantom from a displayed B-scan for a fixed virtual scanner. The task is ill posed: display inver sion supplies magnitude but no phase, while challenge outputs accept only nonnegative scatterer energies. We present phase-pair holographic inversion, a deterministic, training-free two-stage method. Regularized alternating projections (RAP) select a phase favored by the regularized scanner model while restoring the target magnitude after each update. A closed-form phase pair then encodes each complex coefficient as two nonnegative-energy entries half a wavelength apart, matching the central wavelength response and canceling the leading bandwidth error of one shifted scatterer. In a fixed-regularization local-model ablation on all 120 public skin B-scans, each stage improved every scan; their 50-iteration combination reached mean multiscale structural similarity (MS-SSIM) of 0.9935 versus 0.8256 for zero-phase single-scatterer synthesis. On the same scans, organizer-hosted rendering with the final 200-iteration method achieved mean structural MS-SSIM of 0.9943 and median learned per ceptual image patch similarity (LPIPS) of 0.0226. The method provides high-fidelity, scanner-compatible synthesis without recovering the ac quired phase or underlying tissue microstructure. Code is available at https://github.com/MasalaKimchi/SynthOCT-Bibimbap.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/SynthOCT_002.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: https://papers.miccai.org/miccai-2026-sat/supp/SynthOCT_002_supp.pdf
Link to Open Review
Open Review Page: https://openreview.net/forum?id=zbHz9VC3z8
BibTex
@InProceedings{KimJus_PhasePair_MICCAISAT2026,
author = { Kim, Justin Namuk},
title = { { Phase-Pair Holographic Inversion for OCT Digital Phantom Synthesis } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17258},
month = {pending},
page = {pending}
}