Image Matching

Image-Text-Matching (ITM) is one of the defacto methods of learning generalized representations from a large corpus in Vision and Language (VL). However, due to the weak association between the web-collected image-text pairs, models fail to show a fine-grained understanding of the combined semantics of these modalities. To address this issue we propose Hard Negative Captions (HNC): an automatically created dataset containing foiled hard negative captions for ITM training towards achieving fine-grained cross-modal comprehension in VL. Additionally, we provide a challenging manually-created test set for benchmarking models on a fine-grained cross-modal mismatch task with varying levels of compositional complexity. Our results show the effectiveness of training on HNC by improving the models' zero-shot capabilities in detecting mismatches on diagnostic tasks and performing robustly under noisy visual input scenarios. Also, we demonstrate that HNC models yield a comparable or better initialization for fine-tuning

Visual correspondence across image-to-image (2D-2D), image-to-point cloud (2D-3D), and point cloud-to-point cloud (3D-3D) geometric matching forms the foundation for numerous 3D vision tasks. Despite sharing a similar problem structure, current methods use task-specific designs with separate models for each modality combination. We present UniCorrn, the first correspondence model with shared weights that unifies geometric matching across all three tasks. Our key insight is that Transformer attention naturally captures cross-modal feature similarity. We propose a dual-stream decoder that maintains separate appearance and positional feature streams. This design enables end-to-end learning through stack-able layers while supporting flexible query-based correspondence estimation across heterogeneous modalities. Our architecture employs modality-specific backbones followed by shared encoder and decoder components, trained jointly on diverse data combining pseudo point clouds from depth maps with real 3D correspondence annotations. UniCorrn achieves competitive performance on 2D-2D matching and surpasses prior state-of-the-art by 8% on 7Scenes (2D-3D) and 10% on 3DLoMatch (3D-3D) in registration recall. Project website: https://neu-vi.github.io/UniCorrn

Worldwide image geolocalization, which aims to predict the GPS coordinates of any image on Earth, remains challenging due to global visual diversity. Recent generative approaches based on Retrieval-Augmented Generation (RAG) and Large Multimodal Models (LMMs) leverage candidates retrieved from fixed databases for reasoning, but often struggle with scenes that are absent from the reference set. In this work, we propose GeoSearch, an open-world geolocation framework that integrates web-scale reverse image search into the RAG pipeline. GeoSearch augments LMM prompts with database-retrieved coordinates and textual evidence extracted from web pages. To mitigate noise from irrelevant content, we introduce a two-layer filtering mechanism consisting of image matching, followed by confidence-based gating. Experiments on standard benchmarks Im2GPS3k and YFCC4k demonstrate the superiority of GeoSearch under leakage-aware evaluation. Our code and data are publicly available to support reproducibility.

The semantic transferability of general-purpose visual features does not guarantee geometric consistency across images. Using frozen DINOv3 features, we show that geometrically correct correspondences often fall below rank one in cosine similarity yet remain within a small top-$K$ candidate set. This motivates a preserve-then-resolve design: we retain multiple candidates in a many-to-many (m-to-m) association graph and defer their disambiguation to robust estimation. We study m-to-m robust estimation from a probabilistic perspective. We interpret the existing Matching Cardinality Maximization (MCM) mechanism as a dominant-cardinality approximation to likelihood maximization and propose a faster, real-valued mechanism called Harmonic Consensus Maximization (HCM). A two-stage LO-RANSAC uses HCM for candidate sourcing and MCM for graph-aware selection. We evaluate end-task gains in relative-pose estimation, where our pipeline consistently improves rank-one baselines for DINOv2, DINOv3, V-JEPA 2.1, and SigLIP 2. Code is available at https://github.com/LIAS-CUHKSZ/preserve_then_resolve.

Local feature matching plays a critical role in understanding the correspondence between cross-view images. However, traditional methods are constrained by the inherent local nature of feature descriptors, limiting their ability to capture non-local scene information that is essential for accurate cross-view correspondence. In this paper, we introduce SceneGlue, a scene-aware feature matching framework designed to overcome these limitations. SceneGlue leverages a hybridizable matching paradigm that integrates implicit parallel attention and explicit cross-view visibility estimation. The parallel attention mechanism simultaneously exchanges information among local descriptors within and across images, enhancing the scene's global context. To further enrich the scene awareness, we propose the Visibility Transformer, which explicitly categorizes features into visible and invisible regions, providing an understanding of cross-view scene visibility. By combining explicit and implicit scene-level awareness, SceneGlue effectively compensates for the local descriptor constraints. Notably, SceneGlue is trained using only local feature matches, without requiring scene-level groundtruth annotations. This scene-aware approach not only improves accuracy and robustness but also enhances interpretability compared to traditional methods. Extensive experiments on applications such as homography estimation, pose estimation, image matching, and visual localization validate SceneGlue's superior performance. The source code is available at https://github.com/songlin-du/SceneGlue.

Multimodal Large Language Models (MLLMs) have demonstrated strong cross-modal reasoning capabilities, yet their potential for vision-only tasks remains underexplored. We investigate MLLMs as training-free similarity estimators for instance-level image-to-image retrieval. Our approach prompts the model with paired images and converts next-token probabilities into similarity scores, enabling zero-shot re-ranking within large-scale retrieval pipelines. This design avoids specialized architectures and fine-tuning, leveraging the rich visual discrimination learned during multimodal pre-training. We address scalability by combining MLLMs with memory-efficient indexing and top-$k$ candidate re-ranking. Experiments across diverse benchmarks show that MLLMs outperform task-specific re-rankers outside their native domains and exhibit superior robustness to clutter, occlusion, and small objects. Despite strong results, we identify failure modes under severe appearance changes, highlighting opportunities for future research. Our findings position MLLMs as a promising alternative for open-world large-scale image retrieval.

Finding matching keypoints between images is a core problem in 3D computer vision. However, modern matchers struggle with large in-plane rotations. A straightforward mitigation is to learn rotation invariance via data augmentation. However, it remains unclear at which stage rotation invariance should be incorporated. In this paper, we study this in the context of a modern sparse matching pipeline. We perform extensive experiments by training on a large collection of 3D vision datasets and evaluating on popular image matching benchmarks. Surprisingly, we find that incorporating rotation invariance already in the descriptor yields similar performance to handling it in the matcher. However, rotation invariance is achieved earlier in the matcher when it is learned in the descriptor, allowing for a faster rotation-invariant matcher. Further, we find that enforcing rotation invariance does not hurt upright performance when trained at scale. Finally, we study the emergence of rotation invariance through scale and find that increasing the training data size substantially improves generalization to rotated images. We release two matchers robust to in-plane rotations that achieve state-of-the-art performance on e.g. multi-modal (WxBS), extreme (HardMatch), and satellite image matching (SatAst). Code is available at https://github.com/davnords/loma.

Real-world image pairs often exhibit both severe degradations and large viewpoint changes, making image restoration and geometric matching mutually interfering tasks when treated independently. In this work, we propose MatRes, a zero-shot test-time adaptation framework that jointly improves restoration quality and correspondence estimation using only a single low-quality and high-quality image pair. By enforcing conditional similarity at corresponding locations, MatRes updates only lightweight modules while keeping all pretrained components frozen, requiring no offline training or additional supervision. Extensive experiments across diverse combinations show that MatRes yields significant gains in both restoration and geometric alignment compared to using either restoration or matching models alone. MatRes offers a practical and widely applicable solution for real-world scenarios where users commonly capture multiple images of a scene with varying viewpoints and quality, effectively addressing the often-overlooked mutual interference between matching and restoration.

Image matching is a fundamental problem in Computer Vision with direct applications in robotics, remote sensing, and geospatial data analysis. We present an analytical and experimental evaluation of classical local feature-based image matching algorithms on satellite imagery, focusing on the Scale-Invariant Feature Transform (SIFT) and the Oriented FAST and Rotated BRIEF (ORB). Each method is evaluated through a common pipeline: keypoint detection, descriptor extraction, descriptor matching, and geometric verification via RANSAC with homography estimation. Matching quality is assessed using the Inlier Ratio - the fraction of correspondences consistent with the estimated homography. The study uses a manually constructed dataset of GPS-annotated satellite image tiles with intentional overlaps. We examine the impact of the number of extracted keypoints on the resulting Inlier Ratio.

Recent semi-dense image matching methods have achieved remarkable success, but two long-standing issues still impair their performance. At the coarse stage, the over-exclusion issue of their mutual nearest neighbor (MNN) matching layer makes them struggle to handle cases with scale difference between images. To this end, we comprehensively revisit the matching mechanism and make a key observation that the hint concealed in the score matrix can be exploited to indicate the scale ratio. Based on this, we propose a scale-aware matching module which is exceptionally effective but introduces negligible overhead. At the fine stage, we point out that existing methods neglect the local consistency of final matches, which undermines their robustness. To this end, rather than independently predicting the correspondence for each source pixel, we reformulate the fine stage as a cascaded flow refinement problem and introduce a novel gradient loss to encourage local consistency of the flow field. Extensive experiments demonstrate that our novel matching pipeline, with these proposed modifications, achieves robust and accurate matching performance on downstream tasks.

Image Goal Navigation (ImageNav) is evaluated by a coarse success criterion, the agent must stop within 1m of the target, which is sufficient for finding objects but falls short for downstream tasks such as grasping that require precise positioning. We introduce AnyImageNav, a training-free system that pushes ImageNav toward this more demanding setting. Our key insight is that the goal image can be treated as a geometric query: any photo of an object, a hallway, or a room corner can be registered to the agent's observations via dense pixel-level correspondences, enabling recovery of the exact 6-DoF camera pose. Our method realizes this through a semantic-to-geometric cascade: a semantic relevance signal guides exploration and acts as a proximity gate, invoking a 3D multi-view foundation model only when the current view is highly relevant to the goal image; the model then self-certifies its registration in a loop for an accurate recovered pose. Our method sets state-of-the-art navigation success rates on Gibson (93.1%) and HM3D (82.6%), and achieves pose recovery that prior methods do not provide: a position error of 0.27m and heading error of 3.41 degrees on Gibson, and 0.21m / 1.23 degrees on HM3D, a 5-10x improvement over adapted baselines.Our project page: https://yijie21.github.io/ain/

Local feature matching has long been a fundamental component of 3D vision systems such as Structure-from-Motion (SfM), yet progress has lagged behind the rapid advances of modern data-driven approaches. The newer approaches, such as feed-forward reconstruction models, have benefited extensively from scaling dataset sizes, whereas local feature matching models are still only trained on a few mid-sized datasets. In this paper, we revisit local feature matching from a data-driven perspective. In our approach, which we call LoMa, we combine large and diverse data mixtures, modern training recipes, scaled model capacity, and scaled compute, resulting in remarkable gains in performance. Since current standard benchmarks mainly rely on collecting sparse views from successful 3D reconstructions, the evaluation of progress in feature matching has been limited to relatively easy image pairs. To address the resulting saturation of benchmarks, we collect 1000 highly challenging image pairs from internet data into a new dataset called HardMatch. Ground truth correspondences for HardMatch are obtained via manual annotation by the authors. In our extensive benchmarking suite, we find that LoMa makes outstanding progress across the board, outperforming the state-of-the-art method ALIKED+LightGlue by +18.6 mAA on HardMatch, +29.5 mAA on WxBS, +21.4 (1m, 10$^\circ$) on InLoc, +24.2 AUC on RUBIK, and +12.4 mAA on IMC 2022. We release our code and models publicly at https://github.com/davnords/LoMa.

Impact craters are a cornerstone of planetary surface analysis. However, while most deep learning pipelines treat craters solely as a detection problem, critical scientific workflows such as catalog deduplication, cross-observation matching, and morphological analog discovery are inherently retrieval tasks. To address this, we formulate crater analysis as an instance-level image retrieval problem and introduce CraterBench-R, a curated benchmark featuring about 25,000 crater identities with multi-scale gallery views and manually verified queries spanning diverse scales and contexts. Our baseline evaluations across various architectures reveal that self-supervised Vision Transformers (ViTs), particularly those with in-domain pretraining, dominate the task, outperforming generic models with significantly more parameters. Furthermore, we demonstrate that retaining multiple ViT patch tokens for late-interaction matching dramatically improves accuracy over standard single-vector pooling. However, storing all tokens per image is operationally inefficient at a planetary scale. To close this efficiency gap, we propose instance-token aggregation, a scalable, training-free method that selects K seed tokens, assigns the remaining tokens to these seeds via cosine similarity, and aggregates each cluster into a single representative token. This approach yields substantial gains: at K=16, aggregation improves mAP by 17.9 points over raw token selection, and at K=64, it matches the accuracy of using all 196 tokens with significantly less storage. Finally, we demonstrate that a practical two-stage pipeline, with single-vector shortlisting followed by instance-token reranking, recovers 89-94% of the full late-interaction accuracy while searching only a small candidate set. The benchmark is publicly available at hf.co/datasets/jfang/CraterBench-R.

Large-scale instance-level training data is scarce, so models are typically trained on domain-specific datasets. Yet in real-world retrieval, they must handle diverse domains, making generalization to unseen data critical. We introduce ELViS, an image-to-image similarity model that generalizes effectively to unseen domains. Unlike conventional approaches, our model operates in similarity space rather than representation space, promoting cross-domain transfer. It leverages local descriptor correspondences, refines their similarities through an optimal transport step with data-dependent gains that suppress uninformative descriptors, and aggregates strong correspondences via a voting process into an image-level similarity. This design injects strong inductive biases, yielding a simple, efficient, and interpretable model. To assess generalization, we compile a benchmark of eight datasets spanning landmarks, artworks, products, and multi-domain collections, and evaluate ELViS as a re-ranking method. Our experiments show that ELViS outperforms competing methods by a large margin in out-of-domain scenarios and on average, while requiring only a fraction of their computational cost. Code available at: https://github.com/pavelsuma/ELViS/

Establishing consistent correspondences across images is essential for 3D vision tasks such as structure-from-motion (SfM), yet most existing matchers operate in a pairwise manner, often producing fragmented and geometrically inconsistent tracks when their predictions are chained across views. We propose MV-RoMa, a multi-view dense matching model that jointly estimates dense correspondences from a source image to multiple co-visible targets. Specifically, we design an efficient model architecture which avoids high computational cost of full cross-attention for multi-view feature interaction: (i) multi-view encoder that leverages pair-wise matching results as a geometric prior, and (ii) multi-view matching refiner that refines correspondences using pixel-wise attention. Additionally, we propose a post-processing strategy that integrates our model's consistent multi-view correspondences as high-quality tracks for SfM. Across diverse and challenging benchmarks, MV-RoMa produces more reliable correspondences and substantially denser, more accurate 3D reconstructions than existing sparse and dense matching methods. Project page: https://icetea-cv.github.io/mv-roma/.

The deep-learning based image matching networks can now handle significantly larger variations in viewpoints and illuminations while providing matched pairs of pixels with sub-pixel precision. These networks have been trained with ground-based image datasets and, implicitly, their performance is optimized for the pinhole camera geometry. Consequently, you get suboptimal performance when such networks are used to match satellite images since those images are synthesized as a moving satellite camera records one line at a time of the points on the ground. In this paper, we present EpiMask, a semi-dense image matching network for satellite images that (1) Incorporates patch-wise affine approximations to the camera modeling geometry; (2) Uses an epipolar distance-based attention mask to restrict cross-attention to geometrically plausible regions; and (3) That fine-tunes a foundational pretrained image encoder for robust feature extraction. Experiments on the SatDepth dataset demonstrate up to 30% improvement in matching accuracy compared to re-trained ground-based models.

Keypoint matching can be slow and unreliable in challenging conditions such as repetitive textures or wide-baseline views. In such cases, known geometric relations (e.g., the fundamental matrix) can be used to restrict potential correspondences to a narrow epipolar envelope, thereby reducing the search space and improving robustness. These epipolar-guided matching approaches have proved effective in tasks such as SfM; however, most rely on coarse spatial binning, which introduces approximation errors, requires costly post-processing, and may miss valid correspondences. We address these limitations with an exact formulation that performs candidate selection directly in angular space. In our approach, each keypoint is assigned a tolerance circle which, when viewed from the epipole, defines an angular interval. Matching then becomes a 1D angular interval query, solved efficiently in logarithmic time with a segment tree. This guarantees pixel-level tolerance, supports per-keypoint control, and removes unnecessary descriptor comparisons. Extensive evaluation on ETH3D demonstrates noticeable speedups over existing approaches while recovering exact correspondence sets.

Reliable product identification from images is a critical requirement in industrial and commercial applications, particularly in maintenance, procurement, and operational workflows where incorrect matches can lead to costly downstream failures. At the core of such systems lies the visual search component, which must retrieve and rank the exact object instance from large and continuously evolving catalogs under diverse imaging conditions. This report presents a structured benchmark of modern visual embedding models for instance-level image retrieval, with a focus on industrial applications. A curated set of open-source foundation embedding models, proprietary multi-modal embedding systems, and domain-specific vision-only models are evaluated under a unified image-to-image retrieval protocol. The benchmark includes curated datasets, which includes industrial datasets derived from production deployments in Manufacturing, Automotive, DIY, and Retail, as well as established public benchmarks. Evaluation is conducted without post-processing, isolating the retrieval capability of each model. The results provide insight into how well contemporary foundation and unified embedding models transfer to fine-grained instance retrieval tasks, and how they compare to models explicitly trained for industrial applications. By emphasizing realistic constraints, heterogeneous image conditions, and exact instance matching requirements, this benchmark aims to inform both practitioners and researchers about the strengths and limitations of current visual embedding approaches in production-level product identification systems. An interactive companion website presenting the benchmark results, evaluation details, and additional visualizations is available at https://benchmark.nyris.io.

Robust stereo visual-inertial odometry (VIO) remains challenging in low-texture scenes and under abrupt illumination changes, where point features become sparse and unstable, leading to ambiguous association and under-constrained estimation. Line structures offer complementary geometric cues, yet many efficient point-line systems still rely on point-guided line association, which can break down when point support is weak and may lead to biased constraints. We present a stereo point-line VIO system in which line segments are equipped with dedicated deep descriptors and matched using an entropy-regularized optimal transport formulation, which performs global mass-transport assignment and supports unmatched observations under ambiguity, outliers, and partial observations. The proposed descriptor is training-free and is computed by sampling and pooling network feature maps. To improve estimation stability, we analyze the impact of line measurement noise and introduce reliability-adaptive weighting to regulate the influence of line constraints during optimization. Experiments on EuRoC and UMA-VI, together with real-world deployments in low-texture and illumination-challenging environments, demonstrate improved accuracy and robustness over representative baselines while maintaining real-time performance.

Acquiring per-frame annotations for ultrasound videos is costly and requires clinical expertise, limiting learning-based analysis. We study cross-video annotation transfer: propagating user-specified annotations from a labeled ultrasound video to an independently acquired target video with no target-side labels or manual initialization. Video trackers and segmentation propagators rely on temporal continuity and require a prompt in every new sequence, whereas cross-image feature matching and one-shot segmentation estimate correspondences independently, without enforcing coherent deformations or supporting both point and mask annotations. We present Match4Annotate, a test-time framework with three stages. A spatiotemporal implicit feature representation lifts frozen vision foundation-model features into a continuous field over space and time, enabling queries beyond the backbone resolution. A continuous implicit feature flow then aligns the source and target fields under a smooth-deformation prior, estimating correspondence in feature space rather than relying on intensity consistency, which is often violated in ultrasound by speckle and acquisition-dependent appearance. Finally, flow-guided annotation transfer uses the estimated flow as a spatial prior over feature similarity. This formulation unifies sparse point and dense mask transfer and includes unconstrained feature matching and direct flow warping as limiting cases. On four clinical ultrasound datasets spanning echocardiography and musculoskeletal imaging, Match4Annotate achieves state-of-the-art annotation transfer, outperforming dense feature-matching baselines across PCK thresholds and one-shot segmentation methods in Dice score. It also demonstrates bidirectional transfer of left-ventricular annotations across datasets. It requires no task-specific training and adapts to each video in minutes on a single consumer GPU.

Establishing reliable image correspondences is essential for many robotic vision problems. However, existing methods often struggle in challenging scenarios with large viewpoint changes or textureless regions, where incorrect cor- respondences may still receive high similarity scores. This is mainly because conventional models rely solely on fea- ture similarity, lacking an explicit mechanism to estimate the reliability of predicted matches, leading to overconfident errors. To address this issue, we propose SURE, a Semi- dense Uncertainty-REfined matching framework that jointly predicts correspondences and their confidence by modeling both aleatoric and epistemic uncertainties. Our approach in- troduces a novel evidential head for trustworthy coordinate regression, along with a lightweight spatial fusion module that enhances local feature precision with minimal overhead. We evaluated our method on multiple standard benchmarks, where it consistently outperforms existing state-of-the-art semi-dense matching models in both accuracy and efficiency. our code will be available on https://github.com/LSC-ALAN/SURE.

Long-term environmental monitoring requires the ability to reconstruct and align 3D models across repeated site visits separated by months or years. However, existing Structure-from-Motion (SfM) pipelines implicitly assume near-simultaneous image capture and limited appearance change, and therefore fail when applied to long-term monitoring scenarios such as coral reef surveys, where substantial visual and structural change is common. In this paper, we show that the primary limitation of current approaches lies in their reliance on post-hoc alignment of independently reconstructed sessions, which is insufficient under large temporal appearance change. We address this limitation by enforcing cross-session correspondences directly within a joint SfM reconstruction. Our approach combines complementary handcrafted and learned visual features to robustly establish correspondences across large temporal gaps, enabling the reconstruction of a single coherent 3D model from imagery captured years apart, where standard independent and joint SfM pipelines break down. We evaluate our method on long-term coral reef datasets exhibiting significant real-world change, and demonstrate consistent joint reconstruction across sessions in cases where existing methods fail to produce coherent reconstructions. To ensure scalability to large datasets, we further restrict expensive learned feature matching to a small set of likely cross-session image pairs identified via visual place recognition, which reduces computational cost and improves alignment robustness.

Image Copy Detection (ICD) aims to identify manipulated content between image pairs through robust feature representation learning. While self-supervised learning (SSL) has advanced ICD systems, existing view-level contrastive methods struggle with sophisticated edits due to insufficient fine-grained correspondence learning. We address this limitation by exploiting the inherent geometric traceability in edited content through two key innovations. First, we propose PixTrace - a pixel coordinate tracking module that maintains explicit spatial mappings across editing transformations. Second, we introduce CopyNCE, a geometrically-guided contrastive loss that regularizes patch affinity using overlap ratios derived from PixTrace's verified mappings. Our method bridges pixel-level traceability with patch-level similarity learning, suppressing supervision noise in SSL training. Extensive experiments demonstrate not only state-of-the-art performance (88.7% uAP / 83.9% RP90 for matcher, 72.6% uAP / 68.4% RP90 for descriptor on DISC21 dataset) but also better interpretability over existing methods.

Composed Image Retrieval (CIR) is the task of retrieving a target image from a database using a multimodal query, which consists of a reference image and a modification text. The text specifies how to alter the reference image to form a ''mental image'', based on which CIR should find the target image in the database. The fundamental challenge of CIR is that this ''mental image'' is not physically available and is only implicitly defined by the query. The contemporary literature pursues zero-shot methods and uses a Large Multimodal Model (LMM) to generate a textual description for a given multimodal query, and then employs a Vision-Language Model (VLM) for textual-visual matching to search for the target image. In contrast, we address CIR from first principles by directly generating the ''mental image'' for more accurate matching. Particularly, we prompt an LMM to generate a ''mental image'' for a given multimodal query and propose to use this ''mental image'' to search for the target image. As the ''mental image'' has a synthetic-to-real domain gap with real images, we also generate a synthetic counterpart for each real image in the database to facilitate matching. In this sense, our method uses LMM to construct a ``paracosm'', where it matches the multimodal query and database images. Hence, we call this method Paracosm. Notably, Paracosm is a training-free zero-shot CIR method. It significantly outperforms existing zero-shot methods on challenging benchmarks, achieving state-of-the-art performance for zero-shot CIR.

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

Standard attention mechanisms are not well suited to stereo matching. Global attention scales quadratically and provides no explicit matching constraint, while local attention is efficient but loses long-range correspondences. We propose MatchAttention, an attention mechanism that embeds an explicit matching constraint into attention by treating the relative position between a query and its matched key as a learnable component of attention sampling. Centering a small contiguous sampling window on this learnable relative position enforces the matching constraint and supports long-range correspondence at strictly linear attention complexity. A differentiable contiguous attention sampling (CAS) operator enables sub-pixel accuracy, and cascaded MatchAttention blocks iteratively refine the relative positions through residual connections. We instantiate MatchAttention as a hierarchical coarse-to-fine stereo network with two variants. MatchAttentionXL targets accuracy and MatchAttentionRT targets real-time edge inference. MatchAttentionXL achieves state-of-the-art accuracy on Middlebury V3 and top results across KITTI 2012/2015 and ETH3D. MatchAttentionRT runs at 9.3 ms on RTX 4060 Ti and 79.1 ms on Jetson Orin NX 16 GB at 1024 x 512, making it the first stereo model to deliver real-time edge inference without sacrificing zero-shot generalization. The code is available at https://github.com/TingmanYan/MatchAttention.

This paper provides a consistent and extensive evaluation of state-of-the-art filtering and refinement methods on common image matching pipelines. Unlike previous comparisons, the designed benchmark also takes into account the more general, real, and practical cases where camera intrinsics are unavailable. Moreover, a novel and effective strategy combining non-deep traditional computer vision approaches based on planar constraints and cross correlation is presented. Experimental analysis provides several insights for current application design and future research directions. In particular, the choice of a proper evaluation protocol discloses the effective differences within the compared solutions which otherwise would tend to flatten. Moreover, the proposed classical algorithmic approach is competitive with recent deep methods. Besides providing robust baseline using traditional computer vision for the evaluation of deep-based methods, this knowledge is useful to improve and better understand the deep image matching architectures. On one hand, geometry-based filtering is effective in presence of outliers without degrading already robust deep pipelines; on the other hand cross-correlation refinement is valid in the case of corner-like keypoints and allows to not directly discard inaccurate matches by default in deep pipelines but to retain and refine them for achieving a better coverage of the scene.

Medical landmark detection usually requires accurate expert annotations, which are laborious and difficult to scale across anatomical regions. In this work, we study an extreme annotation-efficient setting where only a single annotated template image is available. We propose SGB-Match, a structure-guided coarse-to-fine self-supervised matching framework for one-shot medical landmark detection. The framework first learns dense anatomical correspondence from unlabeled augmented image pairs, and then transfers the landmark definition from the annotated template to each target image through feature matching. Different from standard contrastive correspondence learning, where negative candidates are penalized by a structure-agnostic rule, we introduce a structure-guided bias into the contrastive objective. The bias is constructed from relative distance and edge-aware anatomical cues, and explicitly reweights the negative gradients: nearby structure-relevant candidates are weakly repelled, while distant or structure-irrelevant negatives are strongly suppressed. As a result, the learned feature space better preserves local anatomical structures around template landmarks and reduces confusing responses from repeated textures. We further adopt a global-to-local design, where a global encoder provides coarse landmark localization and a local encoder refines the prediction in a cropped region. Extensive experiments on four 2D radiological landmark datasets demonstrate that SGB-Match achieves strong one-shot performance across both public and newly collected datasets and consistently benefits from both structure-guided bias and two-stage refinement.