FAST: Flow Any Scene Transformer

Yongjian Zhang; Longguang Wang; Zhuo Song; Zhiheng Fu; Liang Lin; Yulan Guo

cs.CVSep 30, 2026

FAST: Flow Any Scene Transformer

Authors: Yongjian Zhang, Longguang Wang, Zhuo Song, Zhiheng Fu, Liang Lin, Yulan Guo

Organizations: School of Electronics and Communication Engineering, the Shenzhen Campus of Sun Yat-sen University, Sun Yat-sen University, Shenzhen, China · Hong Kong Polytechnic University, HKSAR, China · School of Computer Science and Engineering, Sun Yat-sen University, Guangzhou, China

Abstract

Scaling has become a primary driver of progress in language and vision foundation models, yet its role in precise correspondence matching remains underexplored. In this work, we present Flow Any Scene Transformer (FAST), a scalable correspondence model driven by two key insights. First, we reveal that the query-key projections inside single-view vision foundation models encode a coarse yet reusable prior for cross-view matching. Second, reusing these pretrained projections in cross-attention form yields a highly effective initialization for a ViT-based matcher built from a single-view encoder. Guided by these insights, we build FAST upon a vanilla single-view foundation model, utilizing a zero-parameter rewiring strategy to convert selected self-attention layers into cross-attention for cross-view interaction. This design allows ViT-based matchers to scale with advances in single-view foundation models, bypassing the need for a dedicated pair-centric pretraining stage. To fully unlock the scaling potential of this formulation, we assemble a 6-million-pair training corpus for general-purpose dense 2D displacement estimation across diverse co-visible image pairs. Extensive experiments demonstrate that FAST achieves state-of-the-art performance across a wide range of benchmarks, while scaling favorably with both backbone size and training data.

Figures & tables

Fig. 1: Zero-shot emergent matching capacity in DINOv3-Large [ 6 ] . Cross-view attention logits (from source queries to target keys) are extracted at the layer- $th$ block. The emergent matching prior is robust to fast-moving objects (row 1) and significant appearance variations (row 2). Warmer colors denote higher attention affinities.

Fig. 2: Comparison of attention responses at initialization. For a given query in the source view, we show its inter-view attention, including the cross-attention and global self-attention responses on the target view. All responses are directly generated using the Q and K projections from the layer- $th$ attention layer in DINOv3 [ 6 ] .

Fig. 3: The proposed zero-parameter attention rewiring strategy, which converts self-attention to cross-attention.

Fig. 4: FAST consists only of a DINOv3 [ 6 ] backbone and a DPT head [ 36 ] . In a single forward pass, it jointly regresses bidirectional optical flow $\{\mathbf{\hat{\mu}}_{0\rightarrow 1},\mathbf{\hat{\mu}}_{1\rightarrow 0}\}$ , confidence maps $\{\mathbf{\hat{C}}_{0},\mathbf{\hat{C}}_{1}\}$ , and covisable mask $\{\mathbf{\hat{M}}_{0},\mathbf{\hat{M}}_{1}\}$ from an RGB image pair.

Method	Sintel		Spring			KITTI
Method	clean	final	EPE	1PE	F1	F1-all	F1-bg	F1-fg
PWCNet [ 11 ]	3.45	4.60	2.288	82.27	4.889	7.72	7.69	7.88
RAFT [ 12 ]	1.61	2.86	1.476	6.790	3.198	5.10	4.74	6.87
SEA-RAFT (M) [ 3 ]	1.44	2.86	0.363	3.686	1.347	4.64	4.47	5.49
CrocoFlow [ 23 ]	1.09	2.44	0.498	4.565	1.508	3.64	3.18	5.94
FlowFormer++ [ 42 ]	1.07	1.94	-	-	-	4.52	-	-

TABLE I: Quantitative comparison on optical flow benchmarks. All metrics lower are better.

Fig. 5: Analysis of the KITTI performance gap. FAST struggles with large out-of-image foreground motions, while its sharper boundary predictions are not fully reflected by KITTI metrics due to annotation ambiguity around object boundaries.

Methods	KT15	KT12	ETH3D	Midd.
Methods	D1-all	D1-all	1PE-noc	2PE-noc
Selective-IGEV [ 43 ]	4.5	3.2	3.4	7.5
MatchAttention-B [ 44 ]	3.78	-	1.27	2.27
S2M2-XL [ 41 ]	2.97	4.05	0.42	1.0
FoundationStereo [ 7 ]	2.8	2.3	0.5	1.1
FlowFormer++ [ 42 ]	6.12	4.50	4.63	10.99

TABLE II: Zero-shot evaluation on stereo matching datasets.

Method	DTU				TA-WB
Method	epe	1pe	2pe	5pe	epe	1pe	2pe	5pe
PanMatch [ 9 ]	26.69	60.7	44.8	32.7	74.36	69.7	55.8	48.0
UFM [ 25 ]	5.55	55.5	32.9	13.8	12.84	51.4	30.6	17.0
RoMav2 [ 27 ]	4.81	38.8	20.5	9.7	10.73	38.9	18.6	11.5
FAST	4.43	54.0	28.7	10.4	12.03	51.0	28.6	14.7

TABLE III: Dense matching performance. Non-occluded metrics are reported.

Method	WxBS	ScanNet			MegaDepth
Method	@10px	@5 ∘	@10 ∘	@20 ∘	@5 ∘	@10 ∘	@20 ∘
LightGlue [ 55 ]	–	17.8	34.0	52.0	51.0	68.1	80.7
LoFTR [ 54 ]	55.4	22.1	40.8	57.6	52.8	69.2	81.2
DKM [ 56 ]	58.9	29.4	50.7	68.3	60.4	74.9	85.1
RoMa [ 8 ]	80.1	31.8	53.4	70.9	62.6	76.7	86.3
UFM [ 25 ]	42.3	31.3	54.1	72.0	41.5	57.9	72.4

TABLE IV: Relative pose estimation results on WxBS [ 50 ] , ScanNet [ 51 , 53 ] , and MegaDepth [ 52 , 54 ] . We evaluate WxBS with mAA metrics while ScanNet and Megadepth with AUC metrics.

Experiment	Stereo Matching				Optical Flow
	Middlebury		ETH3D		Sintel		KITTI 15
	EPE	2PE	EPE	1PE	clean	final	EPE	F1-all
Scratch	5.94	43.47	0.79	16.70	3.02	4.11	4.84	21.52
DINOv2 [ 26 ]	2.01	16.84	0.37	5.02	1.21	2.28	2.27	8.85
DINOv3 [ 6 ]	1.77	13.18	0.28	3.10	1.06	2.13	2.06	7.33
global SA	2.38	17.52	0.31	4.32	1.42	2.44	2.74	11.52

TABLE V: Zero-shot evaluation for ablation study. We ablate initialization , attention , decoder role , parameter scale , model scale , and data scale across multiple datasets. Underlined values denote the baseline settings, and bold indicates the best performance. ∗ denotes evaluation on the training set. For all metrics, lower is better.

Fig. 6: Last-layer comparison between original and rewired DINOv3-Huge (rewired from layer 15 of total layer 32).

Appendix figures & tables7 assets

Supplementary material from the paper’s appendix.

Appendix

Method	Backbone	Cross-view Interaction	Params.	Decoder
CroCo v2 [ 23 ]	ViT [ 57 ]	ViT-B	114M	DPT
UFM [ 25 ]	DINOv2 [ 26 ]	12 $\times$ Transformer	86M	DPT
RoMa v2 [ 27 ]	DINOv3 [ 6 ]	12 $\times$ Transformer + CNN	90M	DPT + Linear
FAST (Ours)	DINOv3 [ 6 ]	Rewired Transformer	0	DPT

Appendix

TABLE VI: Architecture comparison of representative ViT-based correspondence models. “Params.” denotes parameters newly introduced for cross-view interaction, excluding the pretrained backbone and dense prediction heads.

Fig. 7: Optical flow estimation on Sintel and Spring scenes.

Fig. 8: Stereo matching estimation on Booster scenes.

Fig. 9: Dense matching comparison on the challenging WxBS scenes.

Dataset	Dynamic Scenes	Lables			Res.	(Seq, Stride)	Clips
Dataset	Dynamic Scenes	Flow	Depth	Camera	Res.	(Seq, Stride)	Clips
AutoFlow [ 58 ]	✓	✓	✗	✗	448 $\times$ 576	(2,1)	27295
cvo [ 59 ]	✓	✓	✗	✗	512 $\times$ 512	(2,1)	125466
DynamicReplica [ 60 ]	✓	✓	✓	✓	720 $\times$ 1280	(8,4)	35742
FlyingChairs [ 10 ]	✓	✓	✗	✗	384 $\times$ 512	(2,1)	22872
Infinigen [ 61 ]	✓	✓	✓	✓	720 $\times$ 1280	(2,1)	1910

Appendix

TABLE VII: Overview of the training data collection. The datasets are grouped into three categories: dynamic scenes with moving objects, static rigid scenes with small baselines, and static rigid scenes with wide baselines. For video datasets, (Seq, Stride) denotes that continuous frames are divided into clips of length Seq with a temporal stride of Stride . ∗ indicates that optical flow labels are generated via a depth-to-flow strategy described in Appendix B .

Fig. 10: 3D reconstruction of the ETH3D [ 46 ] courtyard scene (38 views). From left to right: raw COLMAP, FAST COLMAP with COLMAP-computed pairs, and FAST with all possible pairs.

Fig. 11: 3D reconstruction of the T&T [ 78 ] Family scene (152 views). From left to right: raw COLMAP, Depth Anything v3, and FAST COLMAP with COLMAP-computed pairs.

Explore similar work

Jun 2, 2026cs.CV

SAMatcher: Co-Visibility Modeling with Segment Anything for Robust Feature Matching

Reliable correspondence estimation is a fundamental problem in image processing, underpinning applications such as Structure from Motion, visual localization, and image registration. Existing learning-based methods have significantly improved local feature representations, yet most still operate at the pixel or patch level and lack explicit modeling of regions that are jointly visible across views. We propose SAMatcher, a feature matching framework that formulates correspondence estimation through co-visibility modeling. Instead of directly matching local features, SAMatcher first predicts co-visible region masks and bounding boxes as structured priors for correspondence estimation. Built upon the Segment Anything Model (SAM), it introduces a symmetric cross-view interaction mechanism that enables bidirectional feature exchange and cross-view semantic alignment. We further develop a unified supervision scheme that jointly optimizes mask prediction and box localization through mask learning, box regression, and mask-box consistency constraints. Extensive experiments on challenging benchmarks demonstrate substantial improvements over existing matching pipelines, particularly under large viewpoint and scale variations. Our results show that foundation models originally designed for monocular segmentation can be effectively extended to multi-view correspondence reasoning through explicit co-visibility modeling, offering a new perspective on structured representation learning for image matching. Code and project page: https://xupan.top/Projects/samatcher

Xu Pan, Qiyuan Ma, Mingyue Dong +3

State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing (LIESMARS), Wuhan University, Wuhan, P. R. China · National Key Laboratory of Space Target Awareness, Space Engineering University, Beijing, P. R. China

Harder Better Faster Denser Feature Matching Segment Anything Model

Ask PDF

Aug 11, 2026cs.LG

Cross-View Feature Matching: Survey, Benchmarking, and Foundation-Model Perspectives

Cross-view feature matching aims to establish reliable correspondences across images with large viewpoint variations. Over the past decade, the field has evolved from task-specific models toward increasingly unified and generalizable correspondence models, with recent progress further driven by the emergence of vision foundation models (VFMs). Despite these advances, existing studies remain highly diverse in their problem formulations, model architectures, training paradigms, and evaluation protocols, making it difficult to obtain a unified understanding of the field. In this survey, we present a unified review of cross-view feature matching. We first introduce a structured taxonomy covering feature extraction, single-type feature matcher, multi-type feature matcher, VFMs based methods, training strategy and robust estimation, providing a coherent framework for analysis and comparison. We further examine recent advances, distilling key design principles and highlighting the shift toward unified and generalizable correspondence models. We also provide a unified experimental benchmarking of representative state-of-the-art methods under consistent protocols, enabling fair and comprehensive performance comparisons. In addition, we discuss open challenges and future directions, including efficiency, robustness under extreme conditions, and cross-domain generalization. This survey aims to provide a comprehensive and structured reference for understanding the evolution, current landscape, and future development of cross-view feature matching in the era of vision foundation models.

Songlin Du, Xiaoyong Lu, Zeyu Wu +5

School of Automation, Southeast University, Nanjing 210096, China · School of Computer Science and Technology, Tongji University, Shanghai 201804, China · School of Intelligence Science and Technology and the Institute of Artificial Intelligence, University of Science and Technology Beijing, Beijing 100083, China +2

Harder Better Faster Denser Feature Matching Cross-View Vision Foundation Models Foundation Model

Ask PDF

Nov 19, 2025cs.CV

RoMa v2: Harder Better Faster Denser Feature Matching

Dense feature matching aims to estimate all correspondences between two images of a 3D scene and has recently been established as the gold standard due to its high accuracy and robustness. However, existing dense matchers still fail or perform poorly for many hard real-world scenarios, and high-precision models are often slow, limiting their applicability. In this paper, we attack these weaknesses on a wide front through a series of systematic improvements that together yield a significantly better model. In particular, we construct a novel matching architecture and loss, which, combined with a curated diverse training distribution, enables our model to solve many complex matching tasks. We further make training faster through a decoupled two-stage matching-then-refinement pipeline, and at the same time, significantly reduce refinement memory usage through a custom CUDA kernel. Finally, we leverage the recent DINOv3 foundation model along with multiple other insights to make the model more robust and unbiased. In our extensive set of experiments, we show that the resulting novel matcher sets a new state-of-the-art, being significantly more accurate than its predecessors. Code is available at https://github.com/Parskatt/romav2

Johan Edstedt, David Nordström, Yushan Zhang +7

Linköping University · Chalmers University of Technology · University of Amsterdam +1

Harder Better Faster Denser Feature Matching Dense Prediction

Ask PDF