Temporal Self-Distillation: Learning Visual State Tracking in Videos Without Supervision

We introduce S$^3$T (Self-Supervised Self-Distillation over Time), which, to the best of our knowledge, is the first fully self-contained framework for continuous video state tracking. Our method treats temporal sampling density as privileged information, based on the hypothesis that a denser view of the same clip recovers the running state more accurately. This view serves as the teacher, while a sparse-view student with the same weights learns to match its next-token distribution. The model generates its own target, so training requires no labels, separate teacher, or reward signal, and adds no inference cost. On LLaVA-OneVision-2-8B, S$^3$T improves VSTAT accuracy by $+1.74$ as a single model, $+2.38$ with souping, and $+2.70$ with additional vision-encoder adaptation, while prior self-evolving methods leave state tracking largely unchanged. The capability learned from unlabeled synthetic clips transfers to real videos, improving performance by $+7.95$ on VSTAT-YouTube state-tracking questions and $+4.50$ on MVBench Action Count.

P-CORE: Self-Supervised Surface Consistency for Point-Based Neural Editing

Advances in neural rendering have enabled high-fidelity multi-view reconstruction of 3D scenes. However, free-form non-rigid shape editing remains a significant challenge. Point-based neural representations are highly desirable for multi-view reconstruction because they lack fixed connectivity, which does not constrain the learned surface topology to that of the initialization. Yet this same property causes point-based representations to struggle with holes and surface discontinuities under large deformations. To address this, we propose a novel self-supervised method to enable point-based representations to adapt to large deformations without requiring ground truth multi-view images of deformed geometry. The key idea is to generate random deformations and to ensure consistency in the predicted surface before and after deformation. In particular, the surface prediction from the deformed point cloud should be the same as the deformation applied to the surface prediction from the original point cloud. We incorporate our approach into attention-based point representations, which differ from splatting-based point representations in their use of a learned interpolation kernel between points as opposed to a Gaussian kernel around each point. This learned interpolation kernel can learn to adapt to large deformations, without requiring addition or removal of points. We show that our framework significantly enhances its robustness to large deformations. Experiments on synthetic geometry editing benchmarks (Neural Editor, Objaverse) demonstrate that our approach outperforms existing point-based methods in zero-shot editing and significantly reduces artifacts. Furthermore, qualitative results on the DTU and Mip-NeRF 360 datasets demonstrate our method's effectiveness on real-world scenes.

ProSR: Semantic-Prototype-Guided Discrete Modeling for Physically Consistent SAR Super-Resolution

High-resolution Synthetic Aperture Radar (SAR) imagery is critical for precision analysis such as automatic target recognition, yet its acquisition is costly. Although generative image super-resolution (ISR) models offer a promising alternative, current smooth-approximation based diffusion frameworks often struggle to preserve the coherent scattering statistics, causing stochastic structural distortions that are less consistent with real SAR physics. To address this, we propose Semantic Prototype-Guided Super-Resolution (ProSR), reformulating SAR ISR as a semantically-guided discrete token prediction task within a quantized latent space. By mapping signal features to discrete scattering primitives, ProSR preserves the impulsive nature of SAR without over-smoothing. Furthermore, we integrate a Self-Supervised Learning backbone into SAR ISR to extract label-free semantic priors, overcoming label scarcity. Guided by these priors, we introduce Semantic-Aligned Detail Encoding to decouple high-frequency signals into discrete scattering primitives. In parallel, the Semantic Prototype Map Generator explicitly constructs semantic prototype maps, allowing Prototype-Map-Guided Attention to route the information flows within identical categories and mitigate inter-class interference. To validate our approach, we present a large-scale 0.25m resolution benchmark from the Umbra Open Dataset. Experimental results show ProSR achieves superior visual quality while preserving essential scattering characteristics required for practical SAR applications.

CoViT: Instance-Correspondence Contrastive Learning for Vision Transformer

Vision Transformers (ViT) excel in semantic understanding but fail to discriminate between object instances (e.g., identical embeddings for two dogs), limiting their use in instance-level tasks such as object detection and instance segmentation. We propose Contrastive Vision Transformer (CoViT), a self-supervised learning framework that injects instance-awareness into ViT through geometry-guided contrastive learning. CoViT uniquely coordinates ViT's attention maps and embeddings by constructing triplets: (1) Attention-guided masking: Refine multi-head attention via adaptive thresholding and morphological operations to generate instance masks, identifying foreground anchors; (2) Hardest contrastive mining: For each anchor, computing pairwise embedding similarities to select the intra-instance hardest positive (least similar patch within its mask) and inter-instance hardest negative (most similar patch from other instances), with intra-instance regions masked during negative search. These triplets drive a contrastive loss that simultaneously compresses intra-instance variance and expands inter-instance margins, forcing ViT to discern subtle geometric and appearance differences between instances. CoViT consistently achieves stable performance gains of over 2 AP points across multiple instance-level perception tasks by using ViT as backbone architecture. Notably, CoViT requires no extra decoders or labels, demonstrating that a pure ViT can learn instance-aware representations via inherent attention priors and targeted contrastive constraints. Code and models will be released.

Revisiting Cross-View Completion: Self-Supervised Pre-Training via Reconstruction Error Comparison

Self-supervised pre-training via cross-view completion learns strong features for 3D vision from co-visible regions of image pairs. However, the reference view provides little information for reconstructing non-co-visible patches, implicitly yielding a monocular training signal in these regions. We introduce Gekko, which turns this limitation into a useful signal. The relative improvement of the cross-view reconstruction error over a masked-autoencoder error is a self-supervised proxy for co-visibility: large improvements indicate co-visible regions, negligible ones non-co-visible areas. Gekko is a network, trained from scratch, that jointly performs cross-view completion, masked autoencoding, and per-pixel prediction of this relative improvement, providing an additional binocular signal for all masked regions without any ground-truth 3D annotation. Under identical architectures and training data, Gekko consistently outperforms CroCo on zero-shot correspondence estimation, relative pose estimation, and pointmap regression, with up to 6 times higher accuracy at the strictest relative-pose threshold and a 22% drop in end-point error on ETH3D. The extra channel it learns is itself a strong co-visibility detector on unseen scenes, and Gekko's frozen features outperform released cross-view backbones of comparable or larger size. It can also be trained directly from raw videos with a simple stride-based curriculum, removing the cumbersome 3D preprocessing prior methods require while matching models trained on curated data. Code and pre-trained models are publicly available.

Benchmarking Spatial, Spectral, and Self-Supervised Cues for Face Forgery Detection under Realistic Degradation

Face forgery detectors often achieve strong results on controlled benchmarks, but their reliability under realistic image degradations remains limited. This paper presents a standardized benchmark for face forgery detection using the Multi-Dimensional Face Forgery Image (MFFI) dataset and evaluates performance on both clean and degraded test partitions. We compare six model families, including convolutional networks, transformer-based models, and a frozen self-supervised DINOv3 backbone, across spatial, spectral, and hybrid input representations. The results show that clean-set performance is not a reliable indicator of robustness under compression, resizing, and blurring. Xception with RGB obtains the best clean performance, reaching 0.884 mean ROC-AUC, but degrades substantially on the harder partition. In contrast, frozen DINOv3 achieves the strongest degraded-set result, with 0.726 mean ROC-AUC, while training only a linear classification head. The representation analysis indicates that Fourier-domain cues are most useful when combined with RGB information, whereas purely spectral inputs consistently underperform spatial representations. Qualitative attribution maps further suggest that convolutional detectors focus on localized artifacts, while DINOv3 relies on broader facial structure. These findings reinforce the need for degraded evaluation protocols and highlight self-supervised visual representations as a promising direction for robust face forgery detection. Our source code is publicly available at https://github.com/lucasdocunha/FaceForgery-Benchmark/.

ViTAMINS: An Empirical Study of Training Self-Supervised Vision Transformers with Synthetic Hard Negatives

We introduce ViTAMINS, a method that integrates synthetic hard negatives into unsupervised vision transformer pretraining to improve representation quality. Our approach is thoroughly benchmarked on ImageNet and transfer learning, image retrieval, copy detection, and image, video segmentation tasks. Notably, our proposed negatives give rise to emergent properties, where learned representations contain explicit information about the semantic content of an image and serve as excellent classifiers (up to +11.3% over baselines). ViTAMINS achieves these benefits through simple modifications to existing contrastive frameworks and outperforms competing methods while being more resource efficient, e.g., our ViT-B surpasses V-JEPA with ViT-L. Our findings motivate reconsidering contrastive learning as a simpler yet powerful alternative to dominant generative and self-distillation approaches.

Motion-Saliency Complementary Masked Modeling for Point Cloud Video Understanding

Point cloud video representation learning is crucial for 3D dynamic scene understanding. In this paper, we propose MoSaiC, a novel Motion-Saliency Complementary masked modeling framework for self-supervised point cloud video representation learning. MoSaiC couples three components: Curriculum Motion-Saliency Masking (CMSM), which guides the masking process toward motion-salient tokens under a curriculum schedule; Normal-Flow Motion (NFM) modeling, which supervises the local rigid rotation of each token in the Lie algebra so(3) as an explicit geometric motion target; and Cross-view Token Consistency Prediction (CTCP), which enforces consistency between two complementary masked views at the token level. Together, these components allow MoSaiC to effectively capture both appearance and motion dynamics. Extensive experiments on multiple downstream tasks, including action recognition, temporal action segmentation, and point-level semantic segmentation, demonstrate the effectiveness of our approach.

Polis: 3D Self-Supervision at City Scale

Reliable semantic representations derived from city-scale 3D models are increasingly important for urban analysis, infrastructure monitoring, autonomous systems, and heritage conservation. However, urban scenes of large spatial extent captured through aerial surveying differ substantially from the indoor, object-level, and self-driving LiDAR data used to pretrain most 3D self-supervised models. We introduce Polis, to our knowledge the first application of Sketched Isotropic Gaussian Regularization (SIGReg) as an objective for a native point cloud encoder, and evaluate it through a frozen-feature benchmark spanning fourteen city- and building-scale corpora. Polis combines geometrically matched cosine invariance, SIGReg, and VICReg-style anti-collapse terms with a 12.8k-scene outdoor pretraining mixture and gravity-preserving spatial view sampling. Controlled ablations show that this objective outperforms student--teacher architecture alternatives, as well as Polis versions without anti-collapse terms, on the same representative outdoor corpus. On three pretraining-disjoint city datasets, Polis reaches $23.8\%$ mean mIoU versus $16.3\%$ for the next-best encoder under high-capacity frozen probing, and $17.3\%$ versus $16.1\%$ at a matched point and voxel budget. The same city-scale lead holds on datasets whose training sets were seen in pretraining. On localized terrestrial captures with fine-grained facade and streetscape labels, the ranking reverses. Our results show that distributionally-regularized joint embedding architectures can be successful on challenging city-scale 3D scenes, and that transfer improves when self-supervision is designed for the capture geometry and spatial context of this domain while also revealing the limits of this specialization.

LeVJEPA: Efficient & Scalable Video Pretraining without the Heuristics

Video carries the temporal structure of the physical world, yet learning representations from it has remained computationally expensive: prevailing self-supervised methods either prevent representation collapse through architectural asymmetries, coupling an exponential-moving-average target encoder, a stop-gradient, and a capacity-limited predictor, or circumvent it by reconstructing masked content in pixel space. We introduce LeVJEPA, the first video encoder trained under LeJEPA's collapse-free objective, which dispenses with both. A single encoder is trained with an invariance loss over global and local views of a clip, regularized by SIGReg, which excludes collapse with a provable guarantee. The architecture reduces to an encoder and a projector, and the objective to a single hyperparameter. This formulation admits two properties. First, the cost of pretraining is governed by the number of tokens the encoder observes; uniform random token dropping renders this number small while simultaneously improving downstream accuracy. At matched epochs on identical data, LeVJEPA matches or surpasses V-JEPA 2 across ViT-S/B/L at 5.6 to 20.8x less pretraining compute, and at matched total FLOPs it exceeds the strongest video baseline by 7.6 points on ImageNet-1K while remaining competitive on motion-centric benchmarks. Second, since no asymmetry between branches is required, the encoder can be trained with block-causal attention at no measurable accuracy cost: temporal ordering becomes a property of the encoder itself. Against a compute-matched DINOv2 trained on frames of the same videos, LeVJEPA approaches the image-pretrained encoder on appearance-centric evaluation while nearly doubling its motion-centric accuracy. These results indicate that, once its computational overhead is removed, video becomes a viable and in several respects preferable substrate for general-purpose visual pretraining.

DINOcular: Self-Supervised Visuospatial Representations

We introduce a self-supervised framework for learning joint visuospatial representations from RGB-D observations. While modern vision foundation models are trained almost exclusively on RGB images, many embodied systems have access to explicit depth sensing, which provides geometric information that monocular inputs cannot recover. Our method integrates depth-derived geometric priors with a visual backbone through inter-patch and intra-patch fusion, enabling the model to encode both appearance and spatial structure efficiently. The resulting representation shows promising improvements on 3D awareness while preserving semantic transfer: it outperforms prior methods of comparable scale on multiple 3D geometry benchmarks, and remains competitive when probed for standard RGB-D semantic segmentation tasks.

SSMB: Self-Supervised Local Feature Detection under Motion Blur

Keypoint detection under motion blur remains a significant challenge, as blur distorts local image structure and degrades the repeatability of feature localization. Existing approaches either rely on computationally expensive deblur-then-detect pipelines that may introduce restoration artifacts, or learn to regress the image positions of handcrafted keypoints extracted on sharp images, which reflects the assumptions of the handcrafted detector rather than what is truly repeatable under blur. We present SSMB, a deblur-free, self-supervised keypoint detector for motion-blurred images that requires neither handcrafted detectors nor external pseudo-labels. SSMB introduces the Local Discriminability Enhancement (LDE) module, which restores fine-grained local discriminability after global feature mixing. Training is performed in two stages. First, geometric pretraining on synthetic shapes bootstraps spatially discriminative keypoint detection without any external detector, just from the rendered geometry. Second, blur-aware training on real sharp-blur image pairs learns blur-invariant detection through a multi-component self-supervised objective that enforces cross-domain consistency, geometric alignment, and spatial coverage. Extensive evaluations on keypoint detection, image matching, relative pose estimation, and visual localization under motion blur demonstrate that SSMB establishes a new state-of-the-art among sparse keypoint detectors, consistently outperforming both supervised and self-supervised baselines across all tasks. Code, models, and datasets will be publicly available upon paper acceptance.

Hyperspectral Diffusion Equivariant Imaging (HyDiff-EI): A Self-supervised Framework for Hyperspectral Image Inpainting

A novel Hyperspectral diffusion Equivariant Imaging (HyDiff-EI) framework for solving the hyperspectral image (HSI) inpainting problem has been presented here. Unlike conventional diffusion-based methods that rely on large-scale pretraining, HyDiff-EI is a test-time optimization framework that learns directly from a single corrupted HSI acquisition. This makes it flexible for different sensor configurations and particularly well-suited for practical remote sensing scenarios where large annotated hyperspectral datasets are limited. To address the ill-posed nature of unsupervised inpainting, we embed equivariant consistency constraints within the diffusion process. By leveraging the inherent geometric symmetries and intrinsic characteristics of HSIs, HyDiff-EI bridges the gap between generative diffusion modeling and self-consistent physical priors. We empirically show that coupling diffusion modeling with equivariant priors substantially enhances noise robustness and generalizability. Extensive experiments on real-world datasets including Chikusei, Botswana, and EMIT demonstrate that HyDiff-EI offers remarkable inpainting quality over existing self-supervised and diffusion-based algorithms in both noiseless and noisy cases.

Joint-Embedding Prediction of Masked Point Tubes for Self-Supervised Learning on 4D Point Cloud Videos

Self-supervised representation learning for 4D point cloud videos is challenging because annotations are costly and reconstruction-based pretraining can overemphasize low-level geometric details. We propose a JEPA-style framework that learns from unlabeled spatiotemporal point clouds through latent point-tube prediction. Instead of reconstructing raw coordinates, the model masks spatiotemporal regions and predicts their target representations from visible context representations in feature space. To stabilize latent prediction, we incorporate Sketched Isotropic Gaussian Regularization, which encourages non-collapsed embeddings without relying on explicit reconstruction targets. This formulation aims to capture both spatial structure and temporal dynamics while keeping the pretraining objective aligned with downstream semantic recognition. Experiments on action and gesture recognition benchmarks show that the learned representations improve downstream fine-tuning, limited-label learning, and cross-dataset transfer. These results suggest that JEPA-style latent prediction is a promising alternative to reconstruction-centered pretraining for 4D point cloud videos.

ConsensusTAS: Self-Supervised Temporal Action Segmentation for Long-Horizon Construction Videos

Recognizing sequential construction activities is important for collaborative human-robot work; for example, robots are able to understand workers' current and upcoming actions and provide timely tool delivery or physical support. However, despite extensive research on construction worker activity recognition, existing studies have been limited to classifying activity categories, such as climbing, lifting, and walking, instead of recognizing fine-grained activity transitions from long-horizon sequences. Addressing this problem is challenging because annotating action temporal boundaries in long construction videos is time-consuming. In this study, we propose ConsensusTAS, a label-free, self-supervised learning approach to segment continuous video streams into distinct activity phases by exploiting the internal consensus of candidate segmentations. We evaluated our algorithm on three public datasets, where it outperformed state-of-the-art methods, achieving an F1@10 of 73.08 on GTEA, an F1@10 of 64.33 on Breakfast, and an F1@50 of 33.50 on static-camera videos from Assembly101. We also tested it on real-world construction videos, where post-hoc evaluation showed that the model successfully recognized and segmented actions within the composite activity of bricklaying, such as spreading mortar on a brick, placing the brick, pressing, and aligning. Compared with other temporal action segmentation models that require computationally intensive large vision-language models, our method can run on a CPU, which provides practical value for video surveillance and human-robot collaboration on mobile robotic platforms.

BenthicDINO: Physics-Informed Self-Distillation for View-Invariant Side-Scan Sonar Representations

Automated perception in side-scan sonar (SSS) imagery is severely hindered by physical acoustic artifacts, resulting in representations that inextricably mix intrinsic seabed reflectivity with transient viewing geometries. Existing self-supervised learning (SSL) frameworks rely on augmentations designed for natural images, failing to account for acoustic degradation and explicitly enforce view-invariance. To address this gap, we introduce a physics-informed self-distillation framework built upon the DINOv3 architecture utilizing a ConvNeXt-v2-Tiny backbone to maximize data efficiency. The proposed methodology enforces view-invariance through two primary mechanisms: physically motivated augmentations that simulate speckle noise, range-dependent attenuation, and radiometric miscalibration; and a Hilbert-Schmidt Independence Criterion (HSIC) penalty that explicitly decouples learned dense patch features from physical viewing parameters. Furthermore, we propose a dense, hierarchical feature fusion strategy across all four network stages to preserve fine-grained sediment details alongside deep semantic abstractions. Extensive evaluation demonstrates that the framework natively groups complex benthic topographies into stable, noise-free semantic clusters without relying on manual annotations. During supervised downstream tasks on the S3Seg dataset, the fused representations exhibited exceptional data efficiency, achieving 96% of its absolute peak performance using only 10% of the available annotated data, ultimately reaching a mean Intersection over Union (mIoU) of 71.4% and an overall accuracy of 86.5%.

Human-JEPA: A Human-Centric Vision Model that Perceives and Anticipates

Machines that understand humans should perceive the present and anticipate the future. Existing human-centric vision model are pretrained on human images, set the state of the art in static dense perception, so motion and anticipation are out of reach. Here we present Human-JEPA, a human-centric vision model trained on video by anchored forecasting: dense targets are pinned to a frozen copy of the initialization, preventing a silent collapse of dense perception, and block masks are replaced by a pure past-to-future split, avoiding a five-point action tax and a seventeen-point re-identification collapse. Under frozen probes, Human-JEPA leads the pixel-anchored specialists on pose and person re-identification at 2.7 times fewer parameters, conceding high-resolution dense parsing, and its released predictor head is the first that does not degrade anticipation. A single safely adapted model thus serves both halves of understanding humans.

When does fusing hand-crafted knowledge with learned representations pay? A cost-normalized benchmark of stacking, substitution, and interference

Fusing prior knowledge with data-driven learning is attractive where data is scarce, yet no controlled account says when it helps, is redundant, or harms. We benchmark one fixed hand-crafted knowledge source, a pinned bank of Gabor targets injected only during training at $\sim$2\% overhead, against data-driven alternatives (SimCLR, SimSiam, DINO, ImageNet transfer, augmentation, learned teachers) under one frozen recipe with fixed subsets: 13 datasets, 9 backbones, 150 to 1.28M images, 32--224\,px, 2.5M--86M parameters ($\computeCells$ classification configurations over $\computeRuns$ runs, plus segmentation and detection transplants). Across the training-time combinations we measure, three outcomes recur (decision-level fusion differs). Different-\emph{currency} sources can stack: the prior composes with DeiT augmentation on attention backbones and is worth $+26$ points to ViT-B/16 at $224$\,px, $+6.7$ at twice that budget. Same-currency sources substitute: against effective self-supervised pretraining, the combination never usefully exceeds the better single source. Fusing at full strength into an already-informed initialization interferes in proportion to what it carries: ImageNet transfer, $-15$ to $-17$ points, removed by a weaker auxiliary weight. Frozen-feature diagnostics measured on each source alone separate these outcomes retrospectively but do not predict them: a rule built on them calls one of nine unseen pairs. At a practitioner's own label budget, the frozen-feature gain predicts the end-to-end gain to within $0.17$ points across 30 cells and seven datasets; the underlying decomposition, $Δ= G + \readout(\mathrm{base})$, holds in sign on $\auditRate\%$ of testable cells and is called an unseen backbone family's feature gain in advance. The project page is https://amughrabi.github.io/MomentAux.

Frozen CLIP Priors for Robust Self-Supervised Poisson Inverse Problems

Self-supervised learning for imaging inverse problems is increasingly important in photon-limited settings, where acquiring clean ground truth is impractical and reconstruction must remain stable under dataset and acquisition shifts. This challenge is amplified under Poisson noise, whose signal-dependent statistics interact with sampling operators (e.g., CFA mosaicing). Meanwhile, foundation vision encoders trained at web scale offer distortion-invariant, content-related representations that generalize well across domains, suggesting a promising route to build priors that transfer beyond the training distribution without expensive fine-tuning. This paper proposes an ADMM-inspired unrolled plug-and-play solver for Poisson inverse problems that decouples a closed-form data-consistency update from a parameter-efficient prior. The prior is implemented as a lightweight decoder operating on frozen CLIP RN50 dense multi-scale features, adapting foundation representations with less trainable parameters. For self-supervision, the method integrates GR2R measurement-domain re-corruption with an Equivariant Imaging regularizer via virtual acquisitions. Experiments on Poisson CFA demosaicing and deblurring show competitive quality, improved robustness under shifts, and self-supervised performance approaching supervised training.

Far from the Crowd: Scalable Self-Supervised Learning via Geographic Isolation

Self-supervised pretraining on remote sensing imagery typically treats all samples as equally informative, despite large variability in geographic and visual structure. We propose a curriculum learning strategy for self-supervised Earth observation that ranks samples by geographic isolation, a label-free proxy derived entirely from geolocation metadata already present in geospatial datasets, requiring no image decoding, no model feedback, and no manual annotation. Unlike visual complexity proxies, it scales as O(D log D) with dataset size D and is well-defined for both contrastive and reconstructive objectives. We integrate the proposed measure into MoCoV2 and MAE pretraining and evaluate across three downstream tasks from CopernicusBench (BigEarthNet, DFC-2020, LCZ). Our curriculum reaches baseline final-epoch performance using as few as 20% of the training budget (MAE) and at most 40% (MoCo) of the training budget, and improves final downstream performance by up to +5 mAP on BigEarthNet, with gains of 1-5 points across benchmarks, matching visual-complexity curricula while reducing pre-computation cost by more than 140x (4 s vs. 568 s on SSL4EO). A CKA and effective-rank analysis further reveals that curriculum-trained encoders develop higher-dimensional, more uniformly utilized embedding spaces throughout training.

RVLoss: Runoff Vote Loss for Self-Supervised LiDAR Scene Flow Estimation

LiDAR scene flow estimates point-wise motion between two consecutive scans, referred to as the source and target. Leading self-supervised methods typically minimize the Chamfer loss, the nearest neighbor distance between the flow-compensated source and the target. However, nearest-neighbor search does not enforce motion rigidity, often leading to inconsistent flows within object instances. Existing approaches address this issue with additional regularization terms, but flow consistency among points remains limited, especially for large objects. We propose RVLoss, a self-supervised loss that incorporates motion rigidity by design through a runoff vote mechanism. Our key observation is that the point-wise motion, calculated from nearest neighbor search, can often be grouped into a small set of dominant flow candidates by voting (top-k voting). Furthermore, when compensating the source by these candidates, the flow that best represents the underlying rigid motion often yields the highest consensus after a second voting (top-1 voting). Based on this insight, we incorporate the two-stage runoff vote into loss design and create cluster-wise rigid flows and free-form flows as pseudo-labels for self-supervised learning. RVLoss can be seamlessly integrated into existing feedforward architectures. Experiments on the Argoverse2 2026 Challenge show that models trained with RVLoss achieve state-of-the-art performance among self-supervised approaches, outperforming baseline models trained with alternative loss designs by 20%. Moreover, cross-dataset evaluations demonstrate consistent performance improvements across four additional datasets. Code will be released upon acceptance.

Mask What Matters: Saliency-Guided Video Self-Supervised Learning for Autonomous Driving

Video self-supervised learning through masked spatiotemporal prediction has emerged as a promising paradigm for learning feature representations from unlabeled data. However, existing methods typically rely on random masking, which indiscriminately removes regions irrespective of their semantic or temporal relevance. In ego-centric driving videos, this can weaken the pretext signal since safety-critical cues such as pedestrians, vehicles, lane boundaries, and dynamic interactions often occupy only a small portion of the frame, yet are central to downstream perception. We introduce V-JEPA4A, a domain-specialized variant of V-JEPA for autonomous driving that is pre-trained on publicly available driving videos with a novel saliency-driven masking policy. It accounts for semantically and temporally relevant context. The proposed policy preserves and predicts context according to semantic importance and temporal relevance, yielding more informative representation learning while retaining the efficiency of masked prediction. We evaluate the resulting encoders on four driving benchmarks spanning tracking, semantic segmentation, and depth estimation. The results demonstrate that V-JEPA4A reduces identity switches on BDD100k MOT by 25% over V-JEPA with random masking, achieves 73.2 mIoU on Cityscapes, and 3.75 RMSE on KITTI-2015 depth, while incurring only ~14% additional pre-training iteration overhead.

Emergent 3D Instance Segmentation from Self-Supervised Point Transformers

Unsupervised 3D instance segmentation of outdoor LiDAR scans has traditionally relied on handcrafted geometric priors such as density-based clustering, motion cues, or projected 2D detections. In this work, we investigate whether a frozen, self-supervised point transformer already contains the structural information required to isolate object instances without any handcrafted geometric prior. Using this transformer purely as a feature extractor, we probe its internal representations across the SemanticKITTI, nuScenes, and Waymo Perception datasets. Our analysis yields four core insights: (1) the instance signal concentrates in the attention queries and keys rather than in the values or final output features; (2) output features semantically collapse, merging adjacent same-class objects that the queries and keys keep distinct; (3) this instance signal is bimodal in depth, strongest at the shallowest and deepest encoder stages; and (4) this signal is driven predominantly by the rotary position encoding (RoPE), whose removal collapses its advantage. We put these findings into our method TokenGraph3D, a training-free segmenter that groups points via connected components on a key-similarity graph, using neither density-based clustering nor proximity priors. Under identical prior-free conditions, we substantially outperform output-feature baselines, making the emergent 3D instance structure visible.

CrevasseSeg: A Label-Efficient UAV Crevasse Segmentation Framework

Crevasse mapping from uncrewed aerial vehicle (UAV) imagery matters for glaciological research and for field safety in glaciated terrain. Yet, pixel-level annotation of glacier surfaces is costly and requires domain experts. We introduce CrevasseSeg, a framework for binary segmentation over the terminus of Borebreen, Svalbard, comprising 1,938 unlabelled UAV orthomosaic tiles for self-supervised/unsupervised fine-tuning, 24 labelled tiles for validation and 176 labelled tiles for testing. Using CrevasseSeg, we benchmark five self-supervised objectives -- BYOL, a Jensen-Shannon Divergence (JSD) objective, Barlow-Twins, VICReg, and a combined BYOL-JSD objective -- across three architectures: O-Net, O-Net++, and a DINOv3-initialised O-Net. Each configuration is evaluated under two frozen-feature readouts that differ only in the form of their decision boundary: a linear probe and a non-linear XGBoost classifier fit only on the 24 labelled validation images. Our central finding is a consistent inversion between the two readouts: DINOv3 features are the weakest under linear probing but the strongest under a non-linear readout. A UMAP analysis of the learned feature space shows that DINOv3 fragments pixels into many small clusters in which the classes are locally interleaved, whereas the convolutional architectures (O-Net and O-Net++) embed them onto a single class-sorted manifold. Satellite-pretrained DINOv3 improves over natural-image initialisation across objectives, and our label-efficient DINOv3-ViT-L-Sat-O-Net-BYOL-JSD pipeline reaches 75.33 mDSC / 61.28 mIoU, outperforming standard machine learning baselines fit on the same 24 labelled images with the RGB pixel values used as features. We release CrevasseSeg to support label-efficient segmentation research in remote sensing.

Gaussian-JEPA: Joint-Embedding Predictive Learning for 3D Gaussian Splats

3D Gaussian Splatting (3DGS) represents 3D content with anisotropic primitives that jointly encode geometry and appearance. Fixed-budget encoders consume sampled observations of Gaussian assets, so the same object may be observed through different primitive realizations. Existing self-supervised methods mainly reconstruct masked Gaussian attributes, tying supervision to one sampled realization and requiring an input-space decoder. Latent prediction offers an alternative, but its application to Gaussian tokens requires targets that accommodate coupled attributes and heterogeneous spatial support. We introduce Gaussian-JEPA, which predicts representations of held-out Gaussian token blocks from visible context. An online encoder processes the context, while a shared exponential-moving-average encoder supplies stop-gradient features for multi-scale targets. Complementary target projections and feature-space grounding provide latent supervision without reconstructing Gaussian attributes. We evaluate the features under Gaussian resampling, partial observations, and renderable shape completion, together with transfer to part segmentation and object classification. Compared with matched reconstruction pretraining, Gaussian-JEPA is more consistent across resampled inputs, retains more instance information under partial observations, and provides stronger frozen features for Gaussian completion. These results support latent prediction as an effective objective for reusable 3D Gaussian representations. Code is on the project page (https://amazingren.github.io/Gaussian-JEPA/).

Self-Supervised Topologically Invariant Manifold Learning for Railway Image Quality Assessment

Existing blind image quality assessment (BIQA) methods typically rely on synthetic distortions and subjective annotations, limiting generalization in real-world domains. To address this, we propose a fully self-supervised BIQA framework based on topologically invariant manifold learning under boundary constraints, which constructs a stable quality reference without manual labels. The framework generates progressive background dilution scales via repeated random cropping around each target; exploiting the monotonic degradation of target information density across these scales, it establishes a self-constrained quality manifold. A linearized spatial moment projection eliminates geometric distortions from random cropping; then a monotonicity divergence filter prunes background-sensitive evaluators, isolating an elite pool \(\mathcal{M}_{\text{elite}}\). A robust M-estimator with a principal component stabilizer fuses the metrics into an asymptotically efficient pseudo-ground truth \(q_{\text{PGT}}\), contracting variance toward the Cramér-Rao lower bound. Extensive evaluations demonstrate that the elite evaluator pool, distilled from 11 baseline metrics, secures superior zero-shot transferability across standard synthetic and wild benchmarks (CSIQ, LIVEC, LIVE-2). Concurrently, deployments on the CQU Railway Rolling Stock Surveillance Dataset (2,797 images) yield a manifold cosine similarity \(>0.999\) and a 100.0\% survival rate under industrial extreme stresses, robustly validating its cross-paradigm decoupling and topological resilience.

GhostPoint: Self-Supervised Representation Learning by Hallucinating Occluded LiDAR Structure

3D object detection from LiDAR point clouds is a core problem in autonomous driving. Recent advances in self-supervised learning (SSL) enable scalable pretraining and transfers well to per-point tasks such as semantic and panoptic segmentation, but transfer to 3D detection remains weaker. We analyze recent SSL methods and find that most objectives are defined only on measured LiDAR returns from visible surfaces, leaving occluded and unobserved regions unconstrained. This visible-surface bias can be sufficient for point-wise prediction, but 3D detection requires robustness to missing structure. To address this gap, we propose GhostPoint, an SSL framework that hallucinates latent features in local neighborhoods around discovered instances, generated via a novel instance voxel dilation. In GhostPoint, an encoder processes observed returns, and an additional predictor infers neighborhood representations from observed context. In addition to standard encoder-level supervision, we introduce a predictor-level supervision scheme on sampled voxels from generated neighborhoods. Specifically, observed (visible/masked) voxels match teacher-encoder targets, while unobserved voxels match teacher-predictor hallucinations. This design encourages the learned representation to explicitly model structure beyond observed returns. Extensive evaluations on nuScenes and Waymo demonstrate that our method achieves state-of-the-art performance, consistently improving downstream 3D detection, especially under sparse scans and limited labels.

Self-Supervised Visual On-Policy Distillation

Visual on-policy distillation relies heavily on an informative teacher-student asymmetry, through either a larger, stronger teacher or privileged supervision, such as reference answers or ground-truth regions of interest. This raises a fundamental question: where can informative asymmetry come from when nothing privileged is available? We answer this by inverting where the asymmetry comes from. Rather than adding privileged information to the teacher, we subtract information from the student. This asymmetry creates the same effective learning signal for free as a teacher with access to information unavailable to the student, without ground-truth annotations, rewards, or a separate stronger teacher model. Building on this principle, we introduce Self-Supervised Visual On-Policy Distillation (S$^2$VOPD), a simple yet effective method that constructs on-policy learning signals from asymmetric augmented views. S$^2$VOPD distills the teacher's distribution conditioned on the original image on-policy into the student distribution conditioned on a strongly augmented view of the same image. We systematically explore a broad design space of visual augmentations and uncover that (1) asymmetry matters: all four augmentation families improve performance, while symmetric self-distillation degrades it; (2) strength matters: performance peaks at a moderate strength; and (3) the gap must remain task-consistent: augmentations that completely remove the question-relevant evidence can induce large but uninformative discrepancies. Across six fine-grained perception benchmarks, S$^2$VOPD improves Qwen3.5-4B from 70.7% to 77.4%, above all open-source models compared, up to Qwen3-VL at 235B, and surpasses GPT-5.4. While holding training data the same, it recovers 96% of the improvement achieved by methods with privileged information. Website is at https://williamium3000.github.io/s2vopd

A Controlled Study of Self-Supervised Image and Video Pretraining under Limited Resources

Visual foundation models are a cornerstone of image and video understanding but typically require large amounts of data and computation. The current scale required for pretraining visual foundation models may be unsustainable or unnecessary, and significant benefits arise when effective models can be obtained with fewer resources. To better understand how self-supervised learning (SSL) objectives behave under resource constraints, we conduct a controlled study of image and video SSL objectives under matched data, architecture, and compute budgets. We compare contrastive, reconstruction, feature-prediction, and diffusion objectives and evaluate both standalone and jointly trained image-video SSL formulations across a diverse set of image and video understanding tasks. Our results show that DINOv2-style pretraining consistently provides the strongest overall performance under limited resources. Furthermore, combining DINOv2 with video SSL objectives such as VideoMAE substantially improves image classification and segmentation performance, but degrades video tracking and camera-pose estimation performance, revealing an important tradeoff between semantic and geometric representation learning. These findings suggest that combining image and video SSL objectives can be beneficial in resource-limited settings, while highlighting the need for improved methods that better balance semantic, temporal, and geometric supervision.

STAR: A Spatial-Topology Aware Routing Framework for Generalizable 3D Scene Understanding

Constructing a unified 3D scene understanding model has long been hindered by the topological discrepancies across sensor modalities. While applying the Mixture-of-Experts (MoE) architecture is a flexible approach for multi-domain 3D understanding, we observe that conventional feature-only MoE routers may underrepresent local sampling topology under semantic supervision, making expert allocation difficult when semantic consistency coexists with geometric heterogeneity. To overcome this challenge, we propose STAR (Spatial-Topology Aware Routing Framework). Specifically, we introduce a multi-attribute self-supervised pre-training branch, covering topological and textural variations, to anchor cross-domain structural priors. Building upon this, we design a domain-aware expert branch with two mechanisms: Domain-Spatial-Guided Routing (DSR), which captures local topological variations from spatial context, and Entropy-controlled Dynamic Allocation (EDA), which adjusts the number of activated experts according to routing uncertainty. Together, these branches combine stable cross-domain representation learning with adaptive expert allocation. Extensive experiments across various tasks, encompassing both indoor and outdoor scenes, demonstrate the effectiveness of STAR. It achieves 80.1% mIoU on the ScanNet validation set and 77.2% mIoU on S3DIS, consistently improving over strong baselines. Code is available at our project page (https://xmw666.github.io/STAR/).