Component-Aware Structure-Preserving Style Transfer for Satellite Visual Sim2Real Data Construction

• URL: http://arxiv.org/abs/2605.19624v2
• Date: Wed, 20 May 2026 08:44:35 GMT
• Title: Component-Aware Structure-Preserving Style Transfer for Satellite Visual Sim2Real Data Construction
• Abstract summary: This paper presents a structure-preserving style transfer framework for satellite visual synthetic-to-real data construction.<n>The proposed method achieves the lowest image distribution discrepancy, with an FID of 54.32 and a KID of 0.048.
• Score: 7.240627719297504
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: For camera-based satellite visual sensing, Sim2Real data construction requires images that approach real-domain sensor appearance while retaining the annotations inherited from simulation. Real sensor images of satellite targets with reliable pose labels and component-level masks are difficult to acquire at scale, whereas synthetic rendering provides exact geometric annotations but suffers from a visible appearance gap. This paper presents a component-aware structure-preserving style transfer framework for satellite visual synthetic-to-real data construction. The method builds weakly paired real--synthetic samples from calibrated real acquisition, ArUco-based camera-pose measurement, CAD rendering, and component masks. It then extracts part-wise real-domain style codes from unlabeled real images and injects them into corresponding synthetic satellite regions through mask-aligned modulation. To keep the generated images usable for downstream sensor-data supervision, adversarial training is combined with local contrastive consistency, self-regularization, and edge-preserving constraints. Experiments are conducted on 5,000 rendered satellite images and 100 real images captured in a calibrated setup. The real images provide target-domain appearance references and final evaluation images, while the downstream GDRNet pose estimator is trained only on synthetic or translated synthetic images. Compared with representative image-translation baselines, the proposed method achieves the lowest image distribution discrepancy, with an FID of 54.32 and a KID of 0.048. When the translated data are used to train GDRNet in this target-domain adaptation setting, the ADD pass rate improves to 0.260 and the AUC improves to 0.611. These results indicate that component-level appearance transfer can improve annotation-preserving satellite visual Sim2Real data generation in the considered calibrated setup.

Related papers

• Towards a satellite image manipulation and deepfake localization benchmark dataset [0.5774786149181393]
  verifying the authenticity of satellite imagery has become increasingly critical given advances in generative artificial intelligence.<n>The remote sensing community lacks high-quality, fine-grained manipulation datasets suitable for training and evaluating detection and image forensics algorithms.<n>We describe a preliminary dataset construction process and prototype benchmark dataset for satellite image manipulation detection and localization.
  arXiv  Detail & Related papers  (2026-08-05T13:36:42Z)
• High-Resolution Flood Mapping With Sentinel-1 and Sentinel-2 via Misalignment-Robust Cross-Sensor Learning and Generative Despeckling [1.458346800173785]
  This work presents an integrated flood mapping framework that jointly addresses limitations through curated and novel learning strategies.<n>We introduce a new Sentinel-2 (S2) and Sentinel-1 (S1) dataset covering the contiguous United States, featuring pixel-accurate 10 m water masks with emphasis on challenging weather conditions and urban environments that are underrepresented in existing benchmarks.<n>Experiments using UNet and UNet++ architectures demonstrate strong multispectral performance (AUPRC up to 0.956) and statistically significant improvements in SAR flood mapping when using shift-invariant loss and CVAE-based despeckling.
  arXiv  Detail & Related papers  (2026-06-29T16:19:59Z)
• Bridging the Sim-to-Real Gap in Semiconductor Visual Program Synthesis via Input Binarization [4.275696286826178]
  We propose a visual program synthesis framework in which a Vision-Language Model (VLM) converts inspection images into editable Domain-Specific Language () code describing circuit geometries.<n>Because the VLM is trained solely on synthetic DSL-rendered data, a domain gap arises when processing real Scanning Electron Microscope (SEM) images.<n>We bridge this gap with an input binarization strategy that strips SEM-specific texture and noise, letting the model focus on geometric structure.
  arXiv  Detail & Related papers  (2026-06-01T16:06:20Z)
• Physics-informed simulation framework for realistic sonar image generation and statistical validation [1.4610038284393168]
  ACOUSIM is a framework that evaluates the statistical alignment between synthetic and real sonar imagery.<n>A Gazebo-based environment generates sonar-like images by explicitly controlling seabed texture, illumination-driven shadowing, platform altitude, and noise.
  arXiv  Detail & Related papers  (2026-05-19T11:49:11Z)
• SpatialEdit: Benchmarking Fine-Grained Image Spatial Editing [85.49998134609665]
  Image spatial editing performs geometry-driven transformations, allowing precise control over object layout and camera viewpoints.<n>Current models are insufficient for fine-grained spatial manipulations, motivating a dedicated assessment suite.<n>We introduce SpatialEdit-Bench, a complete benchmark that evaluates spatial editing.<n>We construct SpatialEdit-500k, a synthetic dataset generated with a controllable Blender pipeline.
  arXiv  Detail & Related papers  (2026-04-06T17:54:42Z)
• Unifying UAV Cross-View Geo-Localization via 3D Geometric Perception [51.687842983240564]
  Cross-view geo-localization for Unmanned Aerial Vehicles (UAVs) remains challenging due to the severe geometric discrepancy between oblique UAV imagery and satellite maps.<n>We propose a geometry-aware UAV geo-localization framework that explicitly models the 3D scene geometry to unify coarse place recognition and fine-grained pose estimation.<n>Our method significantly outperforms state-of-the-art baselines, achieving robust meter-level localization accuracy and improved generalization in complex urban environments.
  arXiv  Detail & Related papers  (2026-04-02T08:08:41Z)
• EpiMask: Leveraging Epipolar Distance Based Masks in Cross-Attention for Satellite Image Matching [0.0]
  Deep-learning based image matching networks can handle significantly larger variations in viewpoints and illuminations.<n>We present EpiMask, a semi-dense image matching network for satellite images.
  arXiv  Detail & Related papers  (2026-03-23T00:48:58Z)
• Learning Domain-Invariant Representations for Cross-Domain Image Registration via Scene-Appearance Disentanglement [6.677604052097574]
  Image registration under domain shift is a fundamental challenge in computer vision and medical imaging.<n>We propose SAR-Net, a unified framework that addresses this challenge through principled scene-appearance disentanglement.<n>Our method achieves a median relative Target Registration Error (rTRE) of 0.25%, outperforming the state-of-the-art MEVIS method (0.27% rTRE) by 7.4%, with robustness of 99.1%.
  arXiv  Detail & Related papers  (2026-01-12T07:14:11Z)
• Persistent feature reconstruction of resident space objects (RSOs) within inverse synthetic aperture radar (ISAR) images [0.0]
  This work focuses on recognition of external structures by use of sequential feature detection and tracking.<n>ISAR imagery is generated via a metaheuristic simulator capable of modelling encounters for a variety of deployment scenarios.<n>It is shown that the use of feature tracking within sequences with the proposed approach will increase confidence in feature detection and classification.
  arXiv  Detail & Related papers  (2025-12-17T17:24:50Z)
• Coarse-to-Fine Hierarchical Alignment for UAV-based Human Detection using Diffusion Models [14.696438400081114]
  We introduce a three-stage diffusion-based framework designed to transform synthetic data for UAV-based human detection.<n>Cwd explicitly decouples global style and local content domain discrepancies and bridges those gaps using three modules.<n>Our method achieves up to $+14.1$ improvement of mAP50 on Semantic-Drone benchmark.
  arXiv  Detail & Related papers  (2025-12-15T19:57:36Z)
• AerialMegaDepth: Learning Aerial-Ground Reconstruction and View Synthesis [57.249817395828174]
  We propose a scalable framework combining pseudo-synthetic renderings from 3D city-wide meshes with real, ground-level crowd-sourced images.<n>The pseudo-synthetic data simulates a wide range of aerial viewpoints, while the real, crowd-sourced images help improve visual fidelity for ground-level images.<n>Using this hybrid dataset, we fine-tune several state-of-the-art algorithms and achieve significant improvements on real-world, zero-shot aerial-ground tasks.
  arXiv  Detail & Related papers  (2025-04-17T17:57:05Z)
• On quantifying and improving realism of images generated with diffusion [50.37578424163951]
  We propose a metric, called Image Realism Score (IRS), computed from five statistical measures of a given image. IRS is easily usable as a measure to classify a given image as real or fake. We experimentally establish the model- and data-agnostic nature of the proposed IRS by successfully detecting fake images generated by Stable Diffusion Model (SDM), Dalle2, Midjourney and BigGAN. Our efforts have also led to Gen-100 dataset, which provides 1,000 samples for 100 classes generated by four high-quality models.
  arXiv  Detail & Related papers  (2023-09-26T08:32:55Z)
• Self-Attending Task Generative Adversarial Network for Realistic Satellite Image Creation [0.0]
  We introduce a self-attending task generative adversarial network (SATGAN) to augment synthetic high contrast scientific imagery of resident space objects. The architecture is shown to reduce instances of hallucinatory objects or obfuscation of semantic content in context images.
  arXiv  Detail & Related papers  (2021-11-18T00:49:18Z)
• From Synthetic to Real: Image Dehazing Collaborating with Unlabeled Real Data [58.50411487497146]
  We propose a novel image dehazing framework collaborating with unlabeled real data. First, we develop a disentangled image dehazing network (DID-Net), which disentangles the feature representations into three component maps. Then a disentangled-consistency mean-teacher network (DMT-Net) is employed to collaborate unlabeled real data for boosting single image dehazing.
  arXiv  Detail & Related papers  (2021-08-06T04:00:28Z)
• Unsupervised Metric Relocalization Using Transform Consistency Loss [66.19479868638925]
  Training networks to perform metric relocalization traditionally requires accurate image correspondences. We propose a self-supervised solution, which exploits a key insight: localizing a query image within a map should yield the same absolute pose, regardless of the reference image used for registration. We evaluate our framework on synthetic and real-world data, showing our approach outperforms other supervised methods when a limited amount of ground-truth information is available.
  arXiv  Detail & Related papers  (2020-11-01T19:24:27Z)
• Transferring and Regularizing Prediction for Semantic Segmentation [115.88957139226966]
  In this paper, we exploit the intrinsic properties of semantic segmentation to alleviate such problem for model transfer. We present a Regularizer of Prediction Transfer (RPT) that imposes the intrinsic properties as constraints to regularize model transfer in an unsupervised fashion. Extensive experiments are conducted to verify the proposal of RPT on the transfer of models trained on GTA5 and SYNTHIA (synthetic data) to Cityscapes dataset (urban street scenes)
  arXiv  Detail & Related papers  (2020-06-11T16:19:41Z)
• Deep CG2Real: Synthetic-to-Real Translation via Image Disentanglement [78.58603635621591]
  Training an unpaired synthetic-to-real translation network in image space is severely under-constrained. We propose a semi-supervised approach that operates on the disentangled shading and albedo layers of the image. Our two-stage pipeline first learns to predict accurate shading in a supervised fashion using physically-based renderings as targets.
  arXiv  Detail & Related papers  (2020-03-27T21:45:41Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.