arXiv subjects

Computer vision

Explore arXiv computer-vision papers and cs.CV metadata. Search within the index for image segmentation, object detection and visual representation learning.

Follow this topic with RSS

At least 1,711 records · Page 95Linked to original sources

arXiv2026

MeshSplatBench: A Unified Benchmark for Triangle- and Mesh-Based Neural Rendering

Triangle- and mesh-based neural rendering aims to bridge neural scene representations and existing graphics engines (\textit{e.g.}, Unity and Blender) by leveraging triangle primitives compatible with standard rasterization hardware. However, existing methods are developed and evaluated under inconsistent settings, with limited comparison and little investigation into practical graphics engine deployment. This gap significantly hinders the understanding of their real-world usability. To address this issue, we introduce MeshSplatBench, the first benchmark for systematic evaluation of triangle- and mesh-based neural rendering from native rendering to graphics engine deployment. We propose a hierarchical deployment protocol with two options: (1) Standard deployment, using a conventional opaque mesh pipeline with vertex colors and hardware Z-buffering; and (2) Dedicated deployment, incorporating method-specific engine implementations to preserve appearance and compositing properties (e.g., alpha blending). For mesh splatting, we further introduce a structural audit to evaluate the topological and geometric integrity of exported surfaces for downstream graphics applications. Extensive evaluations reveal three key findings: (1) graphics engine deployment introduces noticeable quality degradation across methods, while mesh splatting approaches achieve relatively better robustness under standard deployment; (2) dedicated deployment can preserve most rendering fidelity at the cost of approximately 6-30$\times$ slowdown; and (3) explicit connectivity and shared vertex indexing in current mesh splatting methods remain insufficient to guarantee manifoldness or global connectivity. Our benchmark demonstrates that rasterizability alone does not imply graphics readiness and highlights the importance of evaluating practical engine compatibility. The benchmark and source code will be publicly released.

cs.GR↗

arXiv2026

SimFuse3D: Source-Guided Target Simulation and Confidence-Guided Multi-Stage Localization Reweighting for Cross-Platform 3D Object Detection

Changes in sensor height and viewpoint alter object-level point distributions, making cross-platform LiDAR unsupervised domain adaptation (UDA) difficult. Self-training uses labeled source scans and unlabeled target scans, yet a retained prediction may provide a useful target location while enclosing sparse foreground returns, background clutter, or points inconsistent with the predicted box. We refer to this mismatch as box-point inconsistency. We introduce SimFuse3D, which preserves the target placement and repairs the associated pseudo object using measured geometry from labeled source scans. Object Memory retrieves a similar labeled source instance. Target Simulation places the retrieved source geometry at the target location, aligns its points with the target viewing geometry, and filters the aligned crop to approximate the target observation. Confidence-Guided Multi-Stage Localization Reweighting (CMLR) maps each target pseudo-object confidence score to a bounded weight shared by RPN localization and R-CNN box regression. All components operate only during adaptation, leaving the detector architecture and inference graph unchanged. Across six cross-platform transfers, SimFuse3D consistently outperforms Pi3DET-Net and achieves the best performance among the compared adaptation methods on nearly all metrics. On nuScenes-to-KITTI, it ranks first among the compared adaptation methods with both evaluated detectors.

cs.CV↗

arXiv2026

WSPolypNet: Weakly Supervised Polyp Localization in Colonoscopy Videos

Because dense frame-level annotation of colonoscopy videos is costly, we propose WSPolypNet, a weakly supervised framework for polyp localization using only video-level labels. WSPolypNet employs a 3D convolutional neural network trained with video-level supervision to generate class activation maps (CAMs), which identify candidate polyp regions without requiring frame-level spatial annotations. The CAM-derived localization cues are further enhanced using a multi-view strategy and provided to MedSAM2 as point prompts. MedSAM2 then propagates segmentation masks across the video, refining the coarse localization cues according to polyp boundaries. WSPolypNet achieved CorLoc scores of 47.80%, 43.68%, and 35.01% at IoU thresholds of 0.3, 0.5, and 0.7, respectively, compared with 36.87%, 33.72%, and 27.94% in the single-view setting. For small polyps, the multi-view strategy improved CorLoc@0.5 from 16.01% to 30.97%. The framework also achieved a recall of 94.51%. These results demonstrate the potential of weakly supervised spatiotemporal learning to substantially reduce spatial annotation requirements for polyp localization in colonoscopy videos.

cs.CV↗

arXiv2026

AcFlow: Controlling Text-to-Image Diffusion Transformers via Learned Conditional Activation Flow

Text-to-image diffusion transformers (DiTs) are powerful generators, yet direct prompting provides limited control interface for style intensity and can fail to suppress unwanted concepts. To enable these controls, we introduce AcFlow, an inference-time controller that transports intermediate layer image-token activations through a learned concept-conditioned velocity field while keeping the base DiT frozen. A textual concept description specifies the desired intervention, while the integration horizon provides a continuous control parameter. The field produces token-varying, activation-dependent updates. With parameters shared across concepts within each task family, one field covers over 15,000 style descriptions or over 1,000 suppression concepts, and generalizes to concepts unseen during training without per-concept fitting. On style control, AcFlow achieves the best style--content trade-off among the evaluated baselines in the high-style-alignment regime. At a fixed operating point, AcFlow attains style--content alignment of 0.5365/0.2860, compared with 0.4397/0.2684 for the baseline with the highest style alignment. On concept suppression, AcFlow reduces the fraction of images showing the concept from 95.3%/82.1% to 41.6%/40.5% on held-in/held-out concepts, including cases where deleting them from the prompt fails to remove them. Our analyses support the learned velocity field as an adaptive control mechanism, with update directions varying across tokens and depending on their activation states. Our code is available at https://github.com/Nove1yst/AcFlow.

cs.CV↗

arXiv2026

Decodable but Misrouted: Sparse Features Uncover a Readout Gap in Vision-Language Models for Harmful Meme Detection

When large vision-language models misclassify harmful memes, the failure may reflect missing internal evidence or an inability to route represented evidence to their outputs. We distinguish these cases in Gemma-3 and Qwen3.5 using sparse autoencoders, role-conditioned probes, causal interventions, and recovery experiments across six harmful content benchmarks, with additional Spanish and Hindi-English code-mixed evaluations. Sparse readouts outperform native prediction on all six primary binary tasks: Qwen averages $0.740$ versus $0.432$ for native macro-F1, residual reconstruction reaches $0.486$, and Gemma improves from $0.532$ to $0.714$. These gains measure how accessible the label is to a supervised readout; they do not show that the model's native generation already applies such a decision rule. Under the evaluated scales, Qwen silent-feature ablation is $24-63$ times more probe-sensitive, whereas routed-feature patching on literal yes/no tasks is $16-140$ times more output-sensitive. Native-only threshold calibration explains much, but not all of the gap: on five tasks with matched probe scores, it recovers $69.8$\% of the raw native-to-probe difference, while direct routing adds $0.094$ mean macro-F1 beyond calibrated native scoring. Joint gold-label, probe-KL, and pairwise LoRA supervision improves dedicated FHM prediction, but a gold-only adapter performs better on the shared seven-task mean. A case study of Gemma-3-12B on the Facebook Hateful Memes dataset finds a distributed rank-32 image-prompt interaction, reaching $0.756$ versus $0.685$ native macro-F1. Robustness controls show that the signal is not explained solely by accompanying OCR and depends on paired visual evidence, and that it extends beyond English. In many of the errors we study, the evidence is represented but does not reach the answer; therefore, routing is a common bottleneck in harmful meme classification.

cs.CV↗

arXiv2026

Beyond UV Mapping: Mesh Texture Compression via Surface-Aligned Texture Fields

Mesh texture compression typically relies on 2D UV atlases, whose chart discontinuities and mapping overhead can limit coding efficiency. To tackle this challenge, we introduce TexF, a surface-aligned texture field that organizes texture attributes in sparse voxels derived from the mesh surface. This representation supports high-resolution textures while preserving local 3D correlations for compression and enabling direct surface queries. For bitstream compression, TexF reuses established 3D attribute codecs, with voxel locations reconstructed from the decoded mesh without separate transmission. For GPU-resident compression, we develop 3DNTC, which combines quantized hash features with a lightweight decoder for random-access reconstruction at surface positions. Differentiable rendering enables image-space refinement of both voxel attributes and compressed neural fields. Experiments on the MPEG and AOM mesh compression benchmarks demonstrate improved average rate-distortion performance over representative UV-based methods for both bitstream and GPU-resident compression. 3DNTC also supports real-time rendering.

cs.CV↗

arXiv2026

Which Terrain Is Better? Preference Learning with VLM Prototypes for Off-Road Traversability Ranking

In vision-based off-road navigation, a robot needs to know not only which obstacles to avoid but also which terrain is better. The first is handled by freespace detection or semantic segmentation. The second is usually answered with a traversability score, but no universal ground truth exists for such a score, so perception falls back on a predefined value per semantic class or a freespace confidence. These scores say what a region is, not which region a robot should prefer. We therefore formulate this preference as visual traversability ranking, an ordering of visible terrain that can be supervised by comparisons between two regions. Standard annotations do not label preference, but they imply its direction. We present TravPro, which converts these annotations into ordered region pairs and fits a small readout on frozen vision--language model (VLM) patch tokens to these pairs. The tokens are clustered once into a fixed prototype bank, and the readout learns a preference score per prototype. The readout is then applied to every patch and serves as a teacher that turns sparse comparisons into dense preference pseudo-labels without pixel-wise annotation. An RGB student distills these maps into a dense terrain-preference map together with a non-ground mask that excludes obstacles and background from the ranking. On five unseen domains, TravPro reaches a mean pairwise accuracy of 0.915 against 0.783 for the strongest baseline, producing an ordering sensitive to surface condition that a per-class value cannot represent. The same VLM and the same supervision yield no such ordering when the VLM is prompted and the supervision is used as dense targets; what matters is how they are used.

cs.CV↗

arXiv2026

HDND: Hierarchical Dynamic Neural Decoding for Multilingual Word/Character Retrieval from Non-Invasive Brain Recordings

While deep learning has enabled language decoding from intracranial brain recordings, extending this capability to non-invasive recordings remains an unresolved challenge. Decoding individual words from non-invasive brain recordings is particularly difficult, as word-level neural evidence is weak, temporally distributed, and entangled with acoustic, lexical, and semantic structure. Existing retrieval pipelines often collapse these factors into a single representation, potentially discarding information available at intermediate temporal scales. Here, we introduce Hierarchical Dynamic Neural Decoding (HDND), a hierarchical dynamic decoding framework that treats word decoding as structured refinement rather than flat label retrieval. HDND combines intermediate neural representations, contextual semantic predictions, and, for selected reading conditions, an auxiliary character-form objective. We evaluate HDND across seven electroencephalography (EEG) and magnetoencephalography (MEG) datasets spanning English, Dutch, Mandarin, and Cantonese listening, reading, and reading-aloud conditions. Across the nine-condition word-retrieval benchmark, the proposed HDND yields a higher participant-averaged balanced Top-10 point estimate than the matched contextual word-decoding baseline in every condition and achieves the highest mean among all compared methods in eight of nine conditions. Across the same nine matched conditions, HDND also yields higher token-micro and pooled word-macro Top-10 point estimates in every setting. Sentence retrieval favors HDND in eight of nine conditions, while auditory speech-segment retrieval is mixed across the six listening conditions. These results show that hierarchical residual refinement can improve multilingual word retrieval from heterogeneous non-invasive brain recordings.

cs.CV↗

arXiv2026

SGAM: Shared Gaussian Geometry with Implicit Amplitude Modeling for Scan-Specific 3D Multi-Contrast MRI Reconstruction

Three-dimensional (3D) multi-contrast magnetic resonance imaging (MCMRI) provides rich anatomical and quantitative information but requires long acquisition times, motivating k-space undersampling. However, reconstruction of large volumetric datasets imposes substantial computational and memory demands. To address this challenge, we propose SGAM, a memory-efficient, scan-specific framework for joint full-volume 3D MCMRI reconstruction. SGAM is based on a shared-geometry Gaussian representation in which amplitudes are modeled by a multi-output implicit neural representation (INR) and contrast-specific phases by explicit variables. This design exploits common anatomical structure across contrasts while preserving contrast-specific signal variations. The representation is jointly optimized using only the acquired multi-coil k-space without external training data. Experiments showed that SGAM consistently outperformed the comparison methods across imaging tasks and acceleration factors, with greater improvements under stronger undersampling. SGAM also achieved a favorable balance between reconstruction quality and computational cost, demonstrating its effectiveness for 3D multi-contrast MRI reconstruction.

cs.CV↗

arXiv2026

AxonSynth: Domain-Randomized Synthetic Data for Zero-Shot 3D Axon Segmentation in Light-Sheet Microscopy

Accurate segmentation of axons in 3D microscopy data is important for analyzing white-matter organization, but dense ground truth labels are expensive to obtain. Existing supervised axon segmentation methods rely on target-domain annotations and can be brittle when tissue type, species, modality, or acquisition conditions change. We present AxonSynth, a domain-randomized synthetic-data framework for training 3D axon segmentation models without manually annotated real training volumes. AxonSynth generates dense synthetic axon labels with orientation priors that reflect realistic fiber configurations and renders them with randomized density, contrast, bias fields, blur, and noise. A three-class 3D U-Net is trained to predict background, axon sheath and intra-axonal space. We evaluate zero-shot transfer on 10 held-out light-sheet microscopy (LSM) patches from macaque and human brain samples labeled with one of three axonal markers, comparing against calibrated thresholding and Frangi filtering using overlap, corrected detection, false-positive, and topology metrics. On macaque samples, AxonSynth achieved the best corrected Dice and corrected precision (0.826 and 0.851), compared with 0.765 and 0.754 for thresholding and 0.685 and 0.762 for Frangi. On human samples, corrected Dice was comparable to thresholding (0.857 vs. 0.868), while component-count error decreased from 22,504 to 3,377. Across all held-out patches, AxonSynth reduced component-count error in 10/10 patches and Euler-characteristic error in 8/10. These results show that synthetic-label domain randomization can reduce dependence on manual axon annotation while supporting synthetic-to-real 3D segmentation.

cs.CV↗

arXiv2026

mmHRI: Towards Privacy-Preserving Human-Robot Interaction with Millimeter-Wave Radar

Assistive robots increasingly operate in many human-centered environments and perform various human-robot interaction (HRI) tasks, such as object delivery. However, most existing HRI systems rely on RGB cameras that continuously observe humans to respond to non-verbal commands, such as hand gestures. This raises privacy concerns in privacy- critical environments, such as hospital wards or restaurants, where direct camera observation of humans is restricted. To develop privacy-preserving HRI, we leverage millimeter-wave (mmWave) radar, which can sense human motion through privacy barriers without identifiable imagery. We propose mmHRI, the first multi-modal robot manipulation framework that achieves mmWave radar-guided privacy-preserving HRI. mmHRI introduces two key designs to mitigate the sparsity and temporal inconsistency of radar data in cluttered robot manipulation environments. First, we propose a dual-stream architecture that jointly learns from unfiltered raw radar tensors and radar point clouds to estimate both human actions and 3D poses. To mitigate signal inconsistency, mmHRI further incorporates a memory-based state-space model (MSSM) that retains historical radar features to reduce abrupt changes in pose/action. These estimated human states are then converted into structured textual robot instructions, which control a vision-language-action (VLA) policy for closed-loop robot manipulation and human-aware reactions. Our evaluation covers human action recognition and closed-loop delivery and retrieval. In the privacy-preserving curtain setting, mmHRI achieves 85.09% action-recognition accuracy, outperforming existing radar-based alternatives. Robot trials further demonstrate successful delivery and retrieval under visual occlusion, with stable task performance across unseen subjects, clutter configurations, and environments.

cs.RO↗

arXiv2026

RoXDrive: Closed-Loop Reinforcement Learning for End-to-End Autonomous Driving via Action-Faithful Rollouts

End-to-end autonomous driving policies are commonly trained via imitation learning on logged demonstrations without observing the consequences of their own actions, leading to causal confusion in closed-loop real-world deployment. To address this issue, reinforcement learning (RL) post-training offers a promising alternative by leveraging world models as interactive training environments to enable future scene generation for policy improvement. Nevertheless, existing approaches either rely on reconstruction-based simulators, offering limited counterfactual interaction, or adopt synthetic simulators to enable long-horizon closed-loop interaction at the cost of a substantial sim-to-real gap. Recently, video world models have exhibited the ability to generate realistic multi-step future rollouts but may not faithfully reflect action conditions, resulting in action-vision mismatch. In this paper, we introduce RoXDrive, a plug-and-play closed-loop RL framework that enables reliable policy optimization by identifying action-faithful world-model rollouts, consisting of two stages: 1) Model pre-training: In addition to imitation-based policy pre-training, we devise an Action-Vision Faithfulness Evaluator for inverse dynamics estimation with our geometry-aware auxiliary trajectory supervision, enabling long-horizon assessment of whether visual dynamics faithfully reflect the conditioning ego actions. 2) Action-faithful RL post-training: Agents iteratively interact with world models to form long-horizon scene rollouts, retaining only action-faithful ones for dense safety-aware scoring and scene-level closed-loop RL post-training. Extensive experiments on nuScenes and an in-house dataset with over 130K training scenarios demonstrate consistent gains across planners, reducing safety violations by 27.6% with DiffusionDrive on nuScenes and 33.7% with Qwen3-VL on the internal data.

cs.CV↗

arXiv2026

V2X-WAM: A Cooperative World Action Model for End-to-End Autonomous Driving

Vehicle-infrastructure cooperation can complement onboard sensing with broader and more informative observations of the traffic environment, providing valuable support for end-to-end autonomous driving. However, existing cooperative driving methods mainly exploit roadside information to enhance the representation of the current scene, while the future consequences of prospective driving actions are rarely modeled explicitly. This limits the ability of the planner to anticipate how its decisions may interact with the evolving traffic environment. To address this issue, we propose V2X-WAM, a cooperative world action model that tightly couples cooperative scene understanding, action generation, and future-world reasoning. V2X-WAM constructs a reliability-aware spatiotemporal representation from vehicle- and infrastructure-side observations, while compressing infrastructure information into a compact quantized message for efficient communication. Based on the resulting cooperative representation, a multimodal planner generates prospective trajectories, which explicitly condition future occupancy and dynamic-flow prediction. The predicted world consequences are then fed back to refine the planned trajectory, forming a closed interaction between action and future-world evolution. Experiments on a large-scale real-world cooperative driving dataset demonstrate that V2X-WAM consistently improves planning accuracy and safety over representative end-to-end cooperative driving methods, while achieving stronger future-world prediction and substantially lower communication overhead. Ablation studies further validate the effectiveness of the proposed design.

cs.RO↗

arXiv2026

Honeycomb: Constant-Size Scene Memory Representation for Video World Models

Video world models require persistent scene memory to maintain consistency during long-horizon video generation. Existing spatial memory systems accumulate RGB observations or latent features, causing storage requirements to grow as generation proceeds. We introduce **Honeycomb**, a video world model built on **HexMemory**, a compact low-rank representation that stores scene features in a fixed-size memory comprising six spatial and spatiotemporal planes. A feed-forward writer maps each newly generated video chunk to plane features. As the spatial coverage or temporal range expands, HexMemory warps the existing planes while preserving their dimensions, then integrates new features through confidence-weighted pooling and a learned residual correction. A reader retrieves latent features from HexMemory to condition subsequent video generation. Because the writer processes only observations from the latest chunk, Honeycomb avoids per-scene optimization and repeated processing of the full generation history. Experiments on WorldScore and RealEstate10K demonstrate strong video generation quality and robust consistency when revisiting previously observed regions, while maintaining constant feature-storage requirements throughout generation. Code and additional visualizations are available on our https://jackswl.github.io/honeycomb/.

cs.CV↗

arXiv2026

VIF-Bench: Evaluating Visual Instruction Following in Multi-Reference Image Generation

Recent multimodal image generation models can take multiple images and textual instructions as input, enabling reference-based generation guided not only by text but also by visual instructions such as layouts, arrows, and pose cues. However, existing benchmarks do not evaluate the joint setting in which multiple references must be composed under multiple and heterogeneous visual-instruction images. To address this gap, we introduce VIF-Bench, a benchmark of 1,241 tasks designed to assess the edge of model capabilities in this joint setting by covering: (i) multi-reference generation (up to 7) under multiple heterogeneous visual instructions (up to 6), (ii) cases where reference images can potentially compete with visual instructions (e.g., a strongly posed subject vs. a target pose), and (iii) controlled comparison of visual instructions with text descriptions at different levels of specificity. Using these capabilities, we uncover three findings: (1) models face an adherence-artifact trade-off: once models reach stronger visual instruction adherence, stronger adherence tends to coincide with more instruction artifacts in generated images, (2) visual instruction adherence tends to be lower on tasks whose reference images carry a salient state of the controlled attribute (e.g., a neon-lit subject under a light-direction instruction), most consistently for light and wind, and (3) for models that can understand visual instructions, it is often better to provide visual constraints directly rather than describe them in text; when using text, a moderate level of detail works better than an exhaustive description. VIF-Bench is released as an open benchmark to establish a basis for fair comparison in controllable multi-reference image generation.

cs.CV↗

arXiv2026

PolyOCR-Venus: Unified OCR Foundation Models for Text-Centric Visual Intelligence

Optical Character Recognition (OCR) is evolving from plain-text transcription toward general visual intelligence, requiring models to recognize, localize, and reason over textual information in complex visual environments. However, existing OCR systems often excel at only some tasks and struggle to balance recognition, parsing, and reasoning across scenarios. In this report, we present PolyOCR, a family of unified OCR foundation models of varying scales. PolyOCR combines a shared instruction-following framework with a large-scale data engine that converts heterogeneous visual resources into quality-verified OCR supervision. We introduce Competence-Guided Policy Optimization, which combines verifier-based Group Relative Policy Optimization with on-policy distillation through sample-wise routing based on teacher reliability and the teacher--student competence gap. We also introduce OCRBench v2.1, our revision of OCRBench v2 with manually verified annotation corrections and task-aligned scoring metrics. Extensive experiments across OCRBench v2.1, CC-OCR, in-house KIE Benchmark, OmniDocBench v1.6 and MDPBench demonstrate that PolyOCR achieves state-of-the-art or highly competitive performance.

cs.CV↗

arXiv2026

CogWAM: Aligning Semantic Cognition with World Action Modeling via Event-Driven Interfaces

Robot policies increasingly incorporate semantic reasoning and future-world prediction, yet combining these capabilities does not guarantee that local predictions and actions remain aligned with task progress. We introduce CogWAM, a cognition-guided world-action model that establishes an explicit semantic interface between task reasoning and world-action learning through a persistent Semantic State, which stores completed task events and the active subtask. CogWAM updates this state only when observations indicate semantic transitions, allowing task-level context to persist across multiple action chunks. To bridge semantic context with physical prediction and control, CogWAM employs progress-conditioned WORLD and ACTION queries that selectively extract task-relevant information for future-world prediction and action generation. During training, the Semantic State provides shared task-progress context for both branches, while inference removes the future-prediction branch and directly generates actions from observations and the maintained state. We further introduce semantic training strategies to improve transition learning and closed-loop conditioning. Without additional robot-action pretraining, CogWAM achieves 15.56 / 11.70 % Score/SR on RoboDojo and state-of-the-art performance on BiCoord, while real-world experiments demonstrate closed-loop dual-arm manipulation with 16.4 fewer Semantic State regenerations than step-wise updating.

cs.RO↗

arXiv2026

The Camera Inside the Editor: Reading the Implicit Camera of Image Editors with Painted Calibration Patterns

Instruction-based image editors insert objects, restyle scenes and render new viewpoints, but it is unknown which camera they assume when they paint into a photograph. Asked to cover the floor with a checkerboard, an editor paints projective structure from which classical vanishing-point geometry reads pitch, roll, focal length, yaw and, on renders, the principal point, without any training. Unlike a calibrator such as GeoCalib, which estimates the camera of an image, this isolates the camera under which the editor paints. On 120 rendered cameras with exact ground truth, Qwen-Image-Edit-2511 paints tile edges that meet their vanishing points within 0.26 degrees, and its implicit camera matches the true one to 0.8 degrees in pitch and 6% in focal length, more accurately than GeoCalib except in roll. Asked to draw the horizon or mark a vanishing point instead, the editor fails, so this knowledge is revealed by painting and not by the explicit tasks we tried. The implicit camera has two priors: roll is pulled towards level (slope 0.71), and telephoto perspective towards a default of about 30 mm, which roughly matches the camera the models paint without any scene. For Qwen, the priors do not grow when blur removes four fifths of the line evidence. They are stronger on real photographs, and on NYUv2 a shorter wording of the task removes the difference for roll. On photographs from a 24--240 mm zoom lens the painted perspective grows with only 0.62 of the lens's slope, while GeoCalib and MoGe-2 saturate at about 52 and 42 mm. FLUX.1 Kontext and LongCat-Image-Edit are pulled much harder. Finally, from a level camera a camera-control LoRA executes pose commands at only 50--70% of their strength, and a board painted into its output agrees with the camera it produced.

cs.CV↗