LTP-BIT: Learning Target Priors Before Image Translation for Remote Sensing

Forum topic · 小凯 · 2026-09-01

Summary

This arXiv paper (2608.28517) by Keyan Hu and colleagues introduces LTP-BIT (Learning the Target Priors Before Image Translation), a prior-first paradigm for cross-modal image translation in remote sensing. The authors observe that existing methods jointly learn target-domain priors and cross-modal dependencies from scarce paired data, overlooking an asymmetry: only the latter requires cross-modal correspondence. Formalized via conditional-score and denoising-risk analyses, LTP-BIT decouples the two tasks: it first learns a target-domain generative prior from large-scale unpaired imagery, then learns source-conditioned control with P-DART, a parameter-efficient dual-stream architecture that retains pretrained backbone weights. Controlled experiments show prior matching mainly improves target-domain realism, while instance fidelity depends more on conditional adaptation. LTP-BIT achieves state-of-the-art results on SAR-to-RGB and NIR-to-RGB benchmarks using only 9.81% task-specific parameters, and on QXS-SAROPT it retains near-full-data instance fidelity with just 25% of paired samples.

Paper Overview

Field: Computer Vision (CV) Authors: Keyan Hu, Mingtao Wang, Ziyu Zhou, Tiandong Shi, Haifeng Li, Ji Qi, Chao Tao arXiv: 2608.28517

Abstract

Cross-modal image translation in remote sensing must preserve source-observed content while matching the target-domain distribution. Existing methods jointly learn the target prior and cross-modal dependence from scarce paired data, overlooking a key asymmetry: only the latter intrinsically requires cross-modal correspondence.

Key Points

The authors formalize this distinction through conditional-score and denoising-risk analyses.
They propose LTP-BIT (Learning the Target Priors Before Image Translation), a prior-first paradigm that decouples the two learning tasks.
LTP-BIT first learns a target-domain generative prior from large-scale unpaired imagery, then retains the pretrained backbone weights and learns source-conditioned control through P-DART, a parameter-efficient dual-stream architecture.
Controlled experiments show that prior matching and scaling mainly improve target-domain realism, while instance fidelity depends more strongly on conditional adaptation.
LTP-BIT achieves state-of-the-art performance on SAR-to-RGB and NIR-to-RGB benchmarks using only 9.81% task-specific parameters.
On QXS-SAROPT, it preserves near-full-data instance fidelity with only 25% of paired samples.

---

*Auto-collected on 2026-09-01*