LoRA-Key creates a standalone user-specific Watermark LoRA trained with a latent watermark prior and GOP, attachable via training-free superposition to protect LoRA ownership while preserving quality.
hub Mixed citations
Dreamsim: Learning new dimensions of human visual similarity using synthetic data
Mixed citation behavior. Most common role is background (50%).
26
Pith papers citing it
Background
50% of classified citations
abstract
Current perceptual similarity metrics operate at the level of pixels and patches. These metrics compare images in terms of their low-level colors and textures, but fail to capture mid-level similarities and differences in image layout, object pose, and semantic content. In this paper, we develop a perceptual metric that assesses images holistically. Our first step is to collect a new dataset of human similarity judgments over image pairs that are alike in diverse ways. Critical to this dataset is that judgments are nearly automatic and shared by all observers. To achieve this we use recent text-to-image models to create synthetic pairs that are perturbed along various dimensions. We observe that popular perceptual metrics fall short of explaining our new data, and we introduce a new metric, DreamSim, tuned to better align with human perception. We analyze how our metric is affected by different visual attributes, and find that it focuses heavily on foreground objects and semantic content while also being sensitive to color and layout. Notably, despite being trained on synthetic data, our metric generalizes to real images, giving strong results on retrieval and reconstruction tasks. Furthermore, our metric outperforms both prior learned metrics and recent large vision models on these tasks.
hub tools
citation-role summary
background 2
dataset 2
method 2
citation-polarity summary
representative citing papers
WBench is a benchmark with 289 test cases and 1,058 turns for evaluating interactive world models using 22 automated metrics validated against human judgments.
Foundation models yield less human-interpretable features than supervised vision transformers, with interpretability tied to activation locality and coarse semantic alignment rather than task performance.
Unfine-tuned MLLMs outperform fine-tuned models on remote sensing image captioning when captions are scored by their ability to reconstruct the source image, and a training-free self-correction method achieves SOTA performance.
Video diffusion models can be adapted into permutation-invariant generators for sparse novel view synthesis by treating the problem as video completion and removing temporal order cues.
PromptEvolver recovers high-fidelity natural language prompts for given images by evolving them via genetic algorithm guided by a vision-language model, outperforming prior methods on benchmarks.
ProDiG progressively transforms aerial Gaussian splats into coherent ground-level 3D reconstructions via diffusion guidance and specialized attention modules.
Noise optimization during sampling recovers diversity in mode-collapsed diffusion models while preserving output fidelity.
MLLM representation spaces are dominated by textual semantics that reduce discriminative power for multimodal retrieval; a whitening transformation called ReAlign corrects the geometry and boosts zero-shot performance.
A new data pipeline using real photos, entity removal, and image-to-video models plus a cross-view attention loss enables text-driven generation of actors in reference scenes with improved alignment.
VLM2Vec converts state-of-the-art vision-language models into universal multimodal embedders via contrastive training on the new MMEB benchmark, delivering 10-20% absolute gains over prior models on both in-distribution and out-of-distribution tasks.
GeoEdit introduces a Lift-Manipulate-Render-Denoise pipeline with dual-branch denoising and variance-homogeneous injection for 3D-consistent object editing in single photos.
Helix4D generates high-quality dynamic 4D meshes from videos by extending Trellis2 with sliding-window cross-frame attention anchored on the first frame and a repurposed 4D temporal encoding.
A multi-teacher distillation framework that packs 50 effect LoRAs and fast sampling into a single adapter while aiming to avoid concept interference.
AttriStory adds a benchmark and AttriLoss-based latent optimization to improve faithful rendering of fine-grained attributes such as clothing color and texture in diffusion-model visual storytelling.
RAE v2 reaches gFID 1.06 on ImageNet-256 in 80 epochs by combining multi-layer encoder sums, complementary REPA targets, and free guidance via output reparameterization.
A technique for parametric stylistic control in latent diffusion models learns disentangled directions from synthetic datasets and applies them via guidance composition while preserving semantics.
Coarse-to-fine 1D token sequences in autoregressive models enable stronger test-time search and even training-free text-to-image generation guided by verifiers, outperforming traditional 2D grid tokenization.
GoViG decomposes goal-conditioned navigation instruction generation into visual state prediction and instruction synthesis using an autoregressive multimodal LLM with one-pass and interleaved reasoning, showing gains on a new R2R-Goal dataset.
Slot-MLLM introduces a slot-attention-based object-centric visual tokenizer with Q-Former encoder, diffusion decoder, and residual vector quantization for improved local visual comprehension and generation in multimodal LLMs.
A generative navigation world model that uses sparse anchored rollout with epipolar constraints to reduce perceptual and geometric drift.
FaceCloak learns a lightweight identity-specific cloaking mask from a single image via synthetic face generation and iterative embedding perturbation to evade multiple recognition models.
RealDiffusion uses heat diffusion as a dissipative prior and a region-aware stochastic process inside a training-free physics-informed attention mechanism to improve multi-character coherence while preserving narrative dynamism in sequential image generation.
SyncFix improves 3D reconstructions by synchronizing multi-view latent representations in a diffusion refinement process, generalizing from pair-wise training to arbitrary view counts at inference.
citing papers explorer
-
Capability $\neq$ Interpretability: Human Interpretability of Vision Foundation Models
Foundation models yield less human-interpretable features than supervised vision transformers, with interpretability tied to activation locality and coarse semantic alignment rather than task performance.
-
Generative Giants, Retrieval Weaklings: Why do Multimodal Large Language Models Fail at Multimodal Retrieval?
MLLM representation spaces are dominated by textual semantics that reduce discriminative power for multimodal retrieval; a whitening transformation called ReAlign corrects the geometry and boosts zero-shot performance.
-
Setting the Stage: Text-Driven Scene-Consistent Image Generation
A new data pipeline using real photos, entity removal, and image-to-video models plus a cross-view attention loss enables text-driven generation of actors in reference scenes with improved alignment.
-
VLM2Vec: Training Vision-Language Models for Massive Multimodal Embedding Tasks
VLM2Vec converts state-of-the-art vision-language models into universal multimodal embedders via contrastive training on the new MMEB benchmark, delivering 10-20% absolute gains over prior models on both in-distribution and out-of-distribution tasks.
-
Improved Baselines with Representation Autoencoders
RAE v2 reaches gFID 1.06 on ImageNet-256 in 80 epochs by combining multi-layer encoder sums, complementary REPA targets, and free guidance via output reparameterization.