DataComp-VLM benchmark shows instruction-heavy data mixing outperforms filtering for VLM training, with DCVLM-Baseline achieving 63.6% on 33 tasks for 8B models (+5.4pp over FineVision).
super hub Mixed citations
DINOv2: Learning Robust Visual Features without Supervision
Mixed citation behavior. Most common role is background (44%).
807
Pith papers citing it
Background
44% of classified citations
abstract
The recent breakthroughs in natural language processing for model pretraining on large quantities of data have opened the way for similar foundation models in computer vision. These models could greatly simplify the use of images in any system by producing all-purpose visual features, i.e., features that work across image distributions and tasks without finetuning. This work shows that existing pretraining methods, especially self-supervised methods, can produce such features if trained on enough curated data from diverse sources. We revisit existing approaches and combine different techniques to scale our pretraining in terms of data and model size. Most of the technical contributions aim at accelerating and stabilizing the training at scale. In terms of data, we propose an automatic pipeline to build a dedicated, diverse, and curated image dataset instead of uncurated data, as typically done in the self-supervised literature. In terms of models, we train a ViT model (Dosovitskiy et al., 2020) with 1B parameters and distill it into a series of smaller models that surpass the best available all-purpose features, OpenCLIP (Ilharco et al., 2021) on most of the benchmarks at image and pixel levels.
hub tools
citation-role summary
method 59
background 57
baseline 9
dataset 3
other 1
citation-polarity summary
claims ledger
- abstract The recent breakthroughs in natural language processing for model pretraining on large quantities of data have opened the way for similar foundation models in computer vision. These models could greatly simplify the use of images in any system by producing all-purpose visual features, i.e., features that work across image distributions and tasks without finetuning. This work shows that existing pretraining methods, especially self-supervised methods, can produce such features if trained on enough curated data from diverse sources. We revisit existing approaches and combine different techniques
authors
co-cited works
representative citing papers
X-Palm supplies the first paired multispectral-to-smartphone palmprint dataset with broad real-world variability to support cross-domain biometric authentication.
Every9D-21M supplies 21.8M real-world 9D pose annotations for 700 everyday categories by propagating manual canonical poses through cross-instance alignment in object-centric videos and verifying them multiview.
A multi-view transformer predicts dense perspective fields that feed a geometric optimizer to estimate camera intrinsics and gravity from arbitrary numbers of real-world views.
Rigel3D jointly generates rigged 3D meshes with geometry, skeleton topology, joint positions, and skinning weights using coupled surface and skeleton latent representations for image-conditioned animation-ready asset synthesis.
Image-to-3D models successfully generate harmful geometries in most cases with under 0.3% caught by commercial filters; existing safeguards are weak but a stacked defense cuts harmful outputs to under 1% at 11% false-positive cost.
neuralCAD-Edit benchmark shows even the best foundation model (GPT 5.2) scores 53% lower than human CAD experts in acceptance trials for multimodal-instructed 3D model edits.
The work creates the first dataset and baseline for generating emission textures on 3D objects to reproduce glowing materials from input images.
Molmo VLMs trained on newly collected PixMo open datasets achieve state-of-the-art performance among open-weight models and surpass multiple proprietary VLMs including Claude 3.5 Sonnet and Gemini 1.5 Pro.
MMMU-Pro is a stricter multimodal benchmark that removes text-only solvable questions, augments options, and requires reading text from images, yielding substantially lower model scores of 16.8-26.9%.
Promptbreeder evolves both task prompts and the mutation prompts that improve them using LLMs, outperforming Chain-of-Thought and Plan-and-Solve on arithmetic and commonsense reasoning benchmarks.
Introduces VICIS task and training framework for inferring visual concepts from image sets, with experiments showing better accuracy, diversity, and generalization than standard VLMs on synthetic and ImageNet data.
InvSplat is a feed-forward multi-view model that predicts 3D Gaussians augmented with intrinsic material attributes for inverse rendering and relighting.
The subspace intervention framework reveals that pre-training objectives shape how ViTs encode geometric information in compressible low-rank subspaces, with peak precision at intermediate layers.
A systematic survey unifies presentation, digital injection, and GenAI synthesis attacks on identity documents, audits datasets for a reality gap, identifies SDGI in multimodal models, and reports APCER above 25% for top models on synthetic IDs.
A training-free prototype memory-guided framework for multi-class prenatal ultrasound anomaly classification and localization using few reference images per class, validated on a 9-category multi-center dataset.
EPO is a trackless, edge-map-alignment framework that refines pose estimates from 3D foundation models and matches or exceeds bundle-adjustment performance with substantially lower runtime and memory use.
GEAR jointly trains VQ tokenizer and AR generator end-to-end via dual hard/soft read-out and representation alignment, achieving up to 10x faster ImageNet gFID convergence than LlamaGen-REPA while generalizing across quantizers and to text-to-image.
WarpHammer densifies scene warps with 3D object priors from generative models and fuses pose-unknown auxiliary views via multi-view geometry to enable stable extreme novel view synthesis.
AnyMatch synthesizes large-scale geometrically consistent multi-modal image pairs from single-view images, enabling fine-tuned matching networks to achieve substantial gains on benchmarks.
A new dataset of 220k+ cross-view pairs and a single-stage geometry-aware model GAGeo based on the π³ 3D foundation model outperforms prior methods on object geo-localization with strong generalization and zero-shot ground-to-drone capability.
First complete digital unwrapping and reading of a Herculaneum papyrus scroll (PHerc. 1667) via synchrotron X-ray CT, virtual unrolling, and machine learning.
Uni-Mo generates 7,488 language-annotated quadruped motions via LLM prompts and video diffusion, lifts them to 3D trajectories, and trains policies achieving 96.7% real-robot success on 392 sampled motions.
Constructs G-equivariant ViTs for arbitrary discrete G ≤ O(2), proves H ≤ G implies G-models embed into H-models and single-head equivariant attention realizes all ordinary G-equivariant maps, introduces D6 hexagonal model, and reports preliminary accuracy gains on PatternNet in low-data regimes.
citing papers explorer
-
CAFOSat: A Strongly Annotated Dataset for Infrastructure-Aware CAFO Mapping Using High-Resolution Imagery
CAFOSat is a new strongly annotated remote-sensing dataset for CAFO mapping that uses human-in-the-loop refinement and curated negatives, with benchmarks on CNNs, transformers, and vision-language models plus a synthetic augmentation pipeline.
-
Spatial Transcriptomics-Guided Alignment Enhances Molecular Profiling in Pathology Foundation Model
STAMP uses a curated 1.8M-pair spatial transcriptomics atlas and pathway-informed alignment to augment pathology foundation models for molecular phenotype inference from H&E WSIs.
-
HARP-VLA: Human-Robot Aligned Representation Learning for Vision-Language-Action Model
HARP aligns human-robot visual and latent action representations via paired bridges and unpaired dynamics supervision to boost VLA policy performance on manipulation tasks.
-
VLM3: Vision Language Models Are Native 3D Learners
Standard VLMs achieve expert-level 3D performance on depth estimation, pose estimation, and object understanding via three simple techniques without architecture changes or regression losses.
-
Fairness Beyond Demographics: Optimizing Performance Across Appearance-Based Hidden Cohorts in Medical Imaging
LHCF trains medical image models for fairness by optimizing across latent appearance-based cohorts discovered via clustering, achieving SOTA results on single and multiple demographic attributes without using any demographic labels.
-
Mining Multi-Modality Spatio-Temporal Cues for Video Important Person Identification
Introduces VIP identification task, releases Temporal-VIP dataset, and presents VIP-Net framework that achieves 67.3% accuracy on identifying important persons in videos while providing rationale similarity of 0.63.
-
Deformable Gaussian Occupancy: Decoupling Rigid and Nonrigid Motion with Factorized Distillation
DeGO decouples rigid and nonrigid motion in Gaussian occupancy prediction via factorized 4D distillation from VGGT, reporting SOTA results on Occ3D-NuScenes with 13.5% gains on human-centric cases.
-
Turning Video Models into Generalist Robot Policies
Decouples action-free video world models from embodiment-specific IDMs using Jacobian-based translation to achieve zero-shot cross-embodiment robot policies.
-
Trinity: Unifying Class-Agnostic Terrain and Semantic Segmentation for Unstructured Outdoor Environments by Leveraging Synthetic Data
Trinity is a unified transformer that performs both class-specific semantic segmentation and class-agnostic terrain segmentation, trained on synthetic RUGDSynth data and evaluated on the new EXTerra real-world dataset.
-
Representation-Conditioned Diffusion Models for Guided Training Data Generation
Representation-conditioned diffusion models generate synthetic ImageNet data that trains classifiers to higher top-1 accuracy than class-conditioned generation (+10.76 pp) or real data (+2.0 pp when scaled).
-
Multi-Modal Building Inspection via Perceiver IO Fusion of Satellite and Street-Level Imagery
A Perceiver IO fusion architecture combines satellite and street-level imagery via DINOv2 tokens and RGB-M masking to classify roof attributes on a new dataset of 32,135 buildings across ten countries.
-
Erased but Exploitable: Black-box Embedding-Aware Prompting Against Unlearned Text-to-Image Diffusion Models
BEAP is a black-box embedding-aware prompting attack using LLM-guided search that raises attack success rate over 60% against unlearned diffusion models while keeping prompts undetectable.
-
A Multimodal 3D Foundation Model for Light Sheet Fluorescence Microscopy Enables Few-Shot Segmentation, Classification, and Deblurring
A multimodal 3D foundation model pretrained on LSM volumes via masked reconstruction and image-text alignment enables improved few-shot segmentation, classification, and deblurring.
-
SplitAvatar: One-shot Head Avatar with Autoregressive Gaussian Splitting
SplitAvatar applies an autoregressive graph splitting network with mesh topology extension and gated density control to generate detailed one-shot head avatars via 3D Gaussian Splatting.
-
Learning from Semantic Dictionaries: Discriminative Codebook Contrastive Learning for Unified Visual Representation and Generation
LEASE achieves state-of-the-art unified performance on ImageNet-1K by combining masked token reconstruction and codebook contrast losses in a one-time precomputed discrete token space.
-
Dual Prototype-Conditioned Diffusion Model for Scalable Multi-Class Unsupervised Anomaly Detection in Large Category Spaces
DPDiff-AD conditions a diffusion model on local prototypes (via nearest aggregation) and global prototypes (via optimal transport) to model normality scalably in multi-class anomaly detection, reporting AUROC gains on 160-category data.
-
EvalVerse: Pipeline-Aware and Expert-Calibrated Benchmarking for Professional Cinematic Video Generation
EvalVerse is a pipeline-aware benchmark that distills expert cinematic judgments into VLMs to assess 'goodness' metrics like aesthetics and multi-shot coherence alongside basic prompt adherence.
-
Label-Efficient Dataset Pruning via Semi-Supervised Pseudo-Labeling
SemiPrune uses a small labeled subset and semi-supervised pseudo-labeling to enable supervised dataset pruning methods, achieving state-of-the-art results on domain-specific, image-corrupted, and long-tailed datasets.
-
UfM*: Uncertainty from Motion* for DNN Depth Estimation Using Gaussians
UfM* uses Gaussian mixtures to compute multiview disagreement for uncertainty in depth estimation with single inference per image, reducing energy and memory use.
-
Uncovering the Latent Potential of Deep Intermediate Representations
Introduces LOES, a constructive spectral method to select task-discriminative subspaces from intermediate layer embeddings, and GeoReg for enforcing simplicial class geometry during fine-tuning, with reported gains increasing with model depth across modalities.
-
Cambrian-P: Pose-Grounded Video Understanding
Cambrian-P adds per-frame camera pose tokens and a regression head to video MLLMs, delivering 4.5-6.5% gains on spatial benchmarks, generalization to other video QA tasks, and SOTA streaming pose estimation on ScanNet.
-
Enhancing Gaze Reasoning in Vision Foundation Models for Gaze Following
A method combining head-conditioned local LoRA adaptation and out-of-cone penalty improves gaze reasoning in vision foundation models, yielding state-of-the-art results on GazeFollow and VAT datasets especially for non-salient targets.
-
SADGE: Structure and Appearance Domain Gap Estimation of Synthetic and Real Data
SADGE is a new fused similarity metric combining DINOv3 appearance and MASt3R geometry via constrained bilinear interaction that correlates with downstream synthetic-to-real performance at Pearson r=0.88 across multiple benchmarks.
-
Divide and Contrast: Learning Robust Temporal Features without Augmentation
Di-COT is an unsupervised contrastive method that stochastically partitions time-series windows into overlapping sub-blocks to learn representations without augmentation, reporting SOTA results on classification and transfer tasks across multiple benchmarks while cutting training time.
-
UniT: Unified Geometry Learning with Group Autoregressive Transformer
UniT unifies online and offline 3D geometry perception via a Group Autoregressive Transformer that processes observation groups with anchor-free point map prediction and a scale-adaptive loss.
-
Spatial Gram Alignment for Ultra-High-Resolution Image Synthesis
Spatial Gram Alignment aligns internal self-similarities of LDM features with foundation priors to reconcile global structure and fine details in ultra-high-resolution text-to-image synthesis.
-
Cross-View Splatter: Feed-Forward View Synthesis with Georeferenced Images
A feed-forward model aligns ground and satellite features to predict Gaussian splats for improved novel-view synthesis on georeferenced outdoor scenes.
-
UniRefiner: Teaching Pre-trained ViTs to Self-Dispose Dross via Contrastive Register
UniRefiner uses contrastive registers and a dual alignment objective to remove three categories of spurious tokens from pre-trained ViTs, yielding up to 9.4% mIoU gains on ADE20K and 22% zero-shot segmentation improvements.
-
Selective, Regularized, and Calibrated: Harnessing Vision Foundation Models for Cross-Domain Few-Shot Semantic Segmentation
HERA is a select-regularize-calibrate framework adapting frozen vision foundation models for cross-domain few-shot semantic segmentation via hierarchical layer selection with ETR, prior-guided regularization, and pixelwise adaptive calibration, reporting over 4.1 mIoU gains.
-
What Makes Synthetic Data Effective in Image Segmentation
Dense scene composition and instance fidelity in synthetic diffusion images drive better segmentation performance; SENSE framework exploits this to improve models on Cityscapes, COCO, and ADE20K.
-
PRISM-SLAM: Probabilistic Ray-Grounded Inference for Scale-aware Metric SLAM
PRISM-SLAM adds a Plücker Ray-Distance Factor and dynamic uncertainty gating to a VFM-augmented factor graph to deliver scale-consistent metric SLAM at 30 FPS from monocular RGB.
-
DexHoldem: Playing Texas Hold'em with Dexterous Embodied System
DexHoldem is a new benchmark providing 1,470 teleoperated demonstrations across 14 manipulation primitives, plus standardized tests for dexterous policy execution and agentic perception in a physical Texas Hold'em setting.
-
Resolving Representation Ambiguity in Feedforward Novel View Synthesis Transformer via Semantic-Spatial Decoupling
Decouples semantic and spatial tokens in NVS transformers to resolve representation ambiguity, yielding consistent gains with near-zero added latency.
-
Speech-Guided Multimodal Learning for Vocal Tract Segmentation in Real-Time MRI
A multimodal training pipeline with phonological bounding-box priors and cross-modal contrastive alignment transfers speech supervision to single-modality rtMRI vocal tract segmentation and outperforms prior methods on two datasets.
-
TIGER-FG: Text-Guided Implicit Fine-Grained Grounding for E-commerce Retrieval
TIGER-FG proposes text-guided implicit fine-grained grounding with dual distillation to address modality and granularity asymmetries in image-to-multimodal e-commerce retrieval, reporting Recall@1 gains of 6.1 and 34.4 points on two new benchmarks.
-
Vision Foundation Models as Generalist Tokenizers for Image Generation
VFMTok builds a generalist image tokenizer on frozen VFMs using adaptive quantization and semantic alignment, delivering gFID 1.36 for autoregressive and 1.25 for continuous generation on ImageNet with 3x faster convergence.
-
GeoFlow: Enforcing Implicit Geometric Consistency in Video Generation
GeoFlow adds a geometry-consistency reward based on rigid camera flow and object appearance preservation, integrated via reinforcement fine-tuning to improve geometric coherence in video generation.
-
A More Word-like Image Tokenization for MLLMs
DiVT clusters patch embeddings into coherent semantic units and adapts token count to image complexity, matching or exceeding baselines with fewer visual tokens on multimodal benchmarks.
-
AURORA: Contextual Orthogonalization for Geometric Representation Learning in Healthcare Foundation Models
AURORA is a representation learning framework that uses contextual orthogonalization and relational alignment to create disentangled, geometrically interpretable latent spaces in healthcare foundation models.
-
SegRAG: Training-Free Retrieval-Augmented Semantic Segmentation
SegRAG is a training-free retrieval-augmented framework that extracts class-specific point prompts from a filtered DINOv3 feature bank to boost SAM3 semantic segmentation performance on standard and agricultural benchmarks.
-
Weighted Reverse Convolution for Feature Upsampling
Weighted Reverse Convolution is a spatially adaptive inverse operator for densifying high-level visual descriptors from vision foundation models, using weighted regularization and an FFT closed-form solution to improve dense prediction tasks.
-
Spatial Blindness in Whole-Slide Multiple Instance Learning
Standard MIL models for whole-slide pathology images exhibit spatial blindness under coordinate permutation; ResTopoMIL separates appearance and spatial learning to restore sensitivity and improve classification and survival prediction.
-
The Learnability Gap in Medical Latent Diffusion
Pretrained autoencoders in medical latent diffusion encode discriminative features well for reconstruction but structure their latent spaces in ways that hinder classifier learning, a gap that persists across architectures and is not closed by domain fine-tuning.
-
Beyond Point-Wise Matching: Structural Representation Alignment for Accelerating Diffusion Transformers
sREPA enforces structural consistency in relational geometry of pre-trained vision features to accelerate DiT training and improve generation quality.
-
Metric-Guided Feature Fusion of Visual Foundation Models for Segmentation Tasks
A label-free metric-guided fusion of complementary features from visual foundation models yields consistent gains in dense prediction tasks with improved object semantics and boundary localization.
-
LACE: Latent Visual Representation for Cross-Embodiment Learning
LACE aligns human-robot visual features via semantic distribution matching on corresponding body parts plus Gram loss, yielding 65% better zero-shot policy transfer than baseline DINO.
-
GeoWorld-VLM: Geometry from World Models for Vision-Language Models
GeoWorld-VLM aligns VLM image features with intermediate representations from camera-conditioned world models via fine-tuning only the encoder and projector, yielding ~4% gains on What'sUp and VSR spatial benchmarks across two VLM backbones.
-
Registers Matter for Pixel-Space Diffusion Transformers
Register tokens enhance pixel-space DiT training and output quality via cleaner high-noise feature maps, and a dual-stream design adds further gains with little overhead.
-
Not All Tasks Quantize Equally: Fisher-Guided Quantization for Visual Geometry Transformer
FGQ applies diagonal Fisher information to guide learnable affine transformations in PTQ for multi-task VGGT, yielding up to 39% relative gains over baselines at 4-bit quantization.
-
DiLA: Disentangled Latent Action World Models
DiLA uses content-structure disentanglement driven by predictive bottlenecks to create semantically structured latent actions for high-fidelity video world models.