Peer Revenue Fully funded

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science Computer Vision

cs.CV Watch Computer Vision

Thursday 8 October

Computer Vision

Parametric Trajectory Distillation for Few-Step Video Generation

Lan Feng, Peter Karkus, Maximilian Igl et al. (+6 more)

Video diffusion and flow models require many sequential evaluations, making generation computationally expensive. Few-step distillation reduces this cost but poses a capacity allocation problem: a student must match the teacher's iterative generation with far less sequential computation. Existing trajectory methods ask the student to reproduce teacher transitions that are highly curved at high noise, which can exceed its capacity and degrade fine detail. We introduce Parametric Trajectory Distillation (PTD), which lets the student parameterize teacher trajectory segments as polynomials and learn from teacher guidance along its own predicted path. PTD is designed to let the learned curvature adapt to the backbone's predictive capacity, preserving motion and diversity. The curvature head is used only in training; inference keeps the original backbone architecture. On Wan2.1-14B, four-step PTD sets a new state of the art for trajectory distillation, significantly improving dynamic quality and naturalness over PDD, the best-performing trajectory-only method on this model, under the same training setting. On the 33B audio-video MiniMax-H3, LoRA-trained PTD significantly improves diversity and naturalness over the state-of-the-art LightX2V Turbo. Blinded human votes give PTD 55.1% and 63.4% preference shares against PDD and LightX2V Turbo. Project page: https://alan-lanfeng.github.io/PTD/.
Read the full abstract

All authors Lan Feng, Peter Karkus, Maximilian Igl, Julius Berner, Yuxiao Chen, Shuhan Tan, Alexandre Alahi, Boris Ivanovic, Marco Pavone

Computer Vision

Stop My Dancing! Understanding, Detecting and Attributing Motion-Aware Deepfake Videos

Fazhong Liu, Yan Meng, Tian Dong et al. (+2 more)

Pose-guided diffusion models can now synthesize entire human figures in motion, spawning a new class of deepfakes: Motion Aware Deepfake (MAD) that have already reached hundreds of millions of viewers. To better understand this emerging threat, we construct the first MAD-specific benchmark and measurement framework, containing over 1.5 million frames that mix 1,363 real and 30,122 synthetic videos from six controllable generators, with realistic perturbations and open-world evaluation splits. Then, we dissect MAD and discover that, despite their global coherence, these videos betray faint yet reliable cues: because the model relies on limited input frames for motion synthesis, it must predict and simulate coherent movement at motion boundaries, thereby producing high-frequency artifacts along with model-specific spectral fingerprints. Based on the observations obtained from analysis on dataset, we propose MoDA, the first defense framework tailored to detect and attribute MAD videos. MoDA couples spatial semantics with steganalysis-rich frequency features via cross-domain alignment and multi-scale aggregation, achieving 94.8% in-distribution and 89.1% cross-dataset detection accuracy gains of 10% to 25% over prior work and 91.5% model attribution accuracy. MoDA achieves 81.94% accuracy on 200 clips produced by two unseen commercial MAD platforms, indicating promising zero-shot transfer, and 78.13% detection accuracy on 1,200 unseen MAD video clips (55k frames in total) collected from the open Internet. Under white-box, gray-box, and black-box adaptive attacks, MoDA maintains relatively stable detection and attribution performance while the accuracies of the baselines drop rapidly.
Read the full abstract

All authors Fazhong Liu, Yan Meng, Tian Dong, Guoxing Chen, Haojin Zhu

Computer Vision

Beyond Resolution: Object-to-Image Ratio Mismatch in Instance Retrieval

Boaz Meivar, Ofir Kedem, Amit Edenzon et al. (+2 more)

Visual instance retrieval often fails when the same object appears at different apparent sizes in the query and gallery. We show that the dominant cause is usually not resolution loss but object-to-image (O2I) ratio mismatch: the object occupies different fractions of the two images. On a controlled benchmark of 3,021 Objaverse objects rendered at five camera distances, more than 80% of the cross-distance degradation is attributable to O2I mismatch rather than resolution for 9 of 12 pretrained backbones; multi-scale architectures cut the resolution-only effect to single digits yet remain equally susceptible. The failure is also asymmetric: tight queries retrieve more reliably against wide gallery images than the reverse. Guided by this analysis, query-side scale augmentation and an OWLv2 crop reranker reach state of the art on ILIAS 100M (29.2 mAP@1000 before reranking, 42.0 after) without training or modifying the precomputed gallery index, and a LoRA fine-tune matches the query-side gains at a single forward pass, showing that O2I robustness is learnable.
Read the full abstract

All authors Boaz Meivar, Ofir Kedem, Amit Edenzon, Gal Chechik, Shai Avidan

Computer Vision

Conditional Residual Prediction: Improving Autoregressive Video Diffusion without a Bidirectional Teacher

Bowen Zheng, Zhiguang Liu, Jiarong Ou et al. (+2 more)

Causal video diffusion models generate video autoregressively, which suits streaming, interactive, and long-video generation. Under standard training, however, they often yield lower generation quality than bidirectional models of the same size. Many existing approaches address this gap by initializing from or distilling a pretrained bidirectional teacher. We instead train a causal model from an image-model initialization, with no bidirectional video model at any stage. Because this path requires neither a large bidirectional teacher nor a complex distillation pipeline, it is simpler and more scalable. On this path, we find that a causal model trained on ground-truth history becomes strongly dependent on it, so that at inference errors in its own generated history propagate forward. We hypothesize that much of this dependence is unnecessary, because the current input already determines much of what the history provides. We propose Conditional Residual Prediction (CRP), a simple recipe for reducing a model's reliance on a condition: the model first predicts the target without the condition, and the condition may only add a residual on top of this prediction. Applied to history, CRP makes the model predict each chunk from the present as far as it can and use the past only for what the present cannot supply. In controlled experiments, CRP nearly closes the 6.14-point gap to a bidirectional model trained under the same setup. Scaling this recipe, we train Optica, a 2B-parameter causal video model that autoregressively generates 5-second 480p videos and reaches 82.78 on VBench with only about 15M training videos.
Read the full abstract

All authors Bowen Zheng, Zhiguang Liu, Jiarong Ou, Rui Chen, Tianyang Hu

Computer Vision

SAGE: Sink-Aware Guided Emphasis for Visual Grounding in Vision-Language Decoders

Jeonghyo Song and YoungJoon Yoo

Recent large vision-language models (VLMs) pair a visual encoder with a large language model (LLM) and perform well on diverse image-text tasks, yet their reliability is often limited by decoder attention pathologies that suppress visual evidence and exacerbate hallucinations. In this paper, we revisit visual attention sinks and uncover a structured, layer-dependent behavior: across prompts, early and late decoder layers exhibit prompt-invariant attention collapse onto the same few image regions, which we term PIS (Prompt-Invariant Sinks), whereas mid layers become prompt-conditioned and drive vision-language alignment. This split suggests that treating sinks as a uniform effect is incomplete. Building on this insight, we propose SAGE (Sink-Aware Guided Emphasis), a lightweight intervention that steers decoder attention away from PIS and toward query-dependent regions of interest (ROIs) using token-aligned ROI masks derived from standard vision backbones such as CLIP, ViT, and DINOv3. Evaluated on diverse vision-encoder + decoder-only LLM VLM families, SAGE improves visual grounding, reduces hallucinations, and yields consistent gains across public downstream vision-language benchmarks, including fine-grained visual discrimination settings where localized evidence is crucial, when instantiated with backbone-derived ROI masks.
Read the full abstract

Computer Vision

ProtoSemImage: Image-Valued Prototypes with Deformable Row Alignment for Interpretable Document Classification

Mohammad Zare and Pirooz Shamsinejadbabaki

Prototypes in classification models are almost always vectors, and a vector has no readable form. This paper asks what happens when a prototype is an image. Documents give the question a natural form, because a document can be rendered as a multi-channel image in which every token becomes a pixel, so a class representative can take the same shape and the same channel semantics as the inputs it stands for. ProtoSemImage represents each class by one or more visual archetypes: prototype images in a four-channel HSV space whose channels carry named linguistic factors. A Skip-Gram objective learns that color space end to end through a four-dimensional bottleneck, discourse boundary rows become differentiable typed difference rows, and classification reduces to 2D visual template matching: a deformable row alignment between a document image and the archetype bank, in the spirit of dynamic time warping. Because the match is a spatial pattern comparison rather than a linear readout, the model reports where an input departs from its archetype and along which channel, and a generative head decodes each archetype back into text. The image representation works: it beats an otherwise identical model with vector prototypes in all three paired seeds, by between 4.3 and 11.8 points on a ten-class task. The distance-based matching does not. A diagnostic that keeps the representation fixed and swaps only the classifier recovers the sequence baselines, which locates a 20.6-point shortfall in the matching rather than in the color compression, and a benchmark built so that a pair of documents shares a bag of words and differs only in arrangement confirms the layout-preservation it was designed for. We report both directions, because for a representation whose whole purpose is inspect ability, the failure modes are as informative as the gains.
Read the full abstract

Computer Vision

Learning to Retrieve: Internalizing Memory Retrieval for Video World Models

JiaKui Hu, Tailai Chen, Yuqi Pan et al. (+8 more)

Video world models aim to generate explorable, 3D-consistent scene videos conditioned on camera trajectories. Existing approaches often rely on external memory systems that explicitly retrieve previously observed content to mitigate scene drift during long-horizon generation. However, these auxiliary memory pathways operate outside the model's internal generative dynamics, preventing the model from intrinsically learning when and what historical information should be retrieved. We propose to internalize memory retrieval into the generation process, allowing retrieval to emerge as an intrinsic behavior of the video world model rather than relying on an external memory system. Based on this principle, we introduce Learning-to-Retrieve (L2R), which repurposes the model's persistent internal state as a memory for historical context. A camera-conditioned retrieval gate selectively accesses relevant historical information from this state, determining what to retrieve, while a retrieval trigger determines when to retrieve. We further supervise the trigger with a 3D re-visibility signal, activating retrieval when previously observed content re-enters the current view while otherwise preserving the existing context. Together, these components enable the model to intrinsically acquire memory retrieval behavior and incorporate relevant historical observations into generation without a separate retrieval pathway. Across multiple base models and camera-revisit benchmarks, L2R improves long-term scene consistency while eliminating the need for an external memory bank or 3D conditions. https://jkhu29.github.io/l2r
Read the full abstract

All authors JiaKui Hu, Tailai Chen, Yuqi Pan, Xuerui Qiu, Jialun Liu, Xiao Cao, Zhenxin Zhu, Guang Chen, Hangjun Ye, Bing Wang, Yanye Lu

Computer Vision

DLC: A Metric-Guided Dynamic Loss Controller for Multi-Objective Training

Jaewan Ko and Janghoon Choi

In this paper, we introduce a metric-guided dynamic loss controller (DLC) for multi-objective image restoration. Conventional image restoration pipelines usually train with a fixed weighted combination of multiple losses, without changing the relative importance of fidelity, perceptual similarity, and no-reference quality during optimization. DLC is an architecture- and loss-term-agnostic training-time controller: it does not modify the restoration architecture or introduce new differentiable loss terms, but dynamically reweights the existing training losses. During training, DLC periodically evaluates the current model on a small fixed feedback subset and uses the resulting quality metrics to update the loss-weight vector through an LLM-based controller. Because DLC operates on existing loss terms rather than task-specific architectures, the same controller formulation can be instantiated across diverse image restoration training pipelines. We evaluate DLC on three restoration domains: low-light image enhancement, deraining, and real-world super-resolution, using both reference-based and no-reference quality metrics. Across these settings, DLC considers metric-dependent trade-offs during optimization and guides training toward balanced operating points across fidelity and perceptual quality. The results show that DLC can move models toward more favorable operating points across different restoration domains, supporting its role as a practical plug-in controller for multi-objective image restoration.
Read the full abstract

Computer Vision

EchoDiST: Self-distillation-based joint learning for diffusion-conditioned echocardiographic myocardial motion estimation

Feiyue Qi, Xingyue Wei and Jianwen Luo

Motion estimation in echocardiography is essential for quantitative assessment of cardiac function and myocardial mechanics, but remains challenging due to image artifacts, limited image information, speckle decorrelation, and the scarcity of ground-truth displacement fields. Anatomy-guided approaches can provide structural information, yet often rely on expert-labeled myocardial segmentations. We propose EchoDiST, a framework for unsupervised echocardiographic myocardial motion estimation that integrates self-distillation-based joint learning with a diffusion-conditioned motion estimation network. Here, unsupervised motion estimation refers to learning without ground-truth displacement fields. The self-distillation strategy jointly optimizes anatomical segmentation and myocardial motion estimation under limited anatomical annotations. Diffusion-based conditioning is used during training with stochastic perturbations, while inference requires only a single deterministic forward pass without iterative reverse-diffusion sampling. EchoDiST was evaluated on three echocardiographic datasets, including two external test datasets under cross-view and cross-dataset settings. Compared with seven representative learning-based methods, EchoDiST consistently improved anatomical alignment, myocardial strain assessment, and motion-derived functional and cardiac-phase assessment. These gains were statistically significant across the evaluated tasks and datasets. Overall, EchoDiST provides an effective approach for reliable myocardial motion estimation under limited anatomical supervision and supports downstream quantitative assessment of cardiac function.
Read the full abstract

Computer Vision

Missing Modality-Aware Calibration for Trustworthy Brain Tumor Segmentation

Sol Lee, Hyunji Kim, Sungrae Hong et al. (+2 more)

Multimodal brain tumor segmentation typically leverages multiple MRI modalities, yet incomplete modality acquisition is common in clinical practice due to protocol heterogeneity and scan failures. Although recent methods maintain segmentation accuracy under missing modality conditions, they frequently overlook prediction reliability, leading to miscalibrated confidence estimates that hinder clinical adoption. Existing calibration techniques are largely modality-agnostic or assume that prediction difficulty decreases monotonically as additional modalities become available. However, in brain tumor segmentation, prediction difficulty depends primarily on which modalities are absent rather than how many, leading to combination-specific and spatially heterogeneous calibration errors. To address this, we propose Missing Modality-Aware Local Temperature Scaling (MMA-LTS), a post-hoc voxel-wise confidence calibration method. It estimates a spatially adaptive temperature field conditioned on a modality-availability learnable token and a voxel-wise difficulty score. Experiments on BraTS 2020 and FeTS 2024 show that MMA-LTS improves calibration while preserving the segmentation accuracy of state-of-the-art models across diverse missing-modality scenarios, thereby enhancing trustworthiness toward clinical deployment.
Read the full abstract

All authors Sol Lee, Hyunji Kim, Sungrae Hong, Donghee Han, Mun Yi

Computer Vision

Fresco++: Frequency-Guided and Canonical-Consistent Optimization for Fine-Grained Head Avatar Modeling

Shikun Zhang, Yong Li, Yiqun Wang et al. (+2 more)

We propose Fresco++, a unified optimization framework for fine-grained and view-consistent head avatar reconstruction. Head avatar optimization is typically driven by per-view image supervision, which can lead to premature fitting of unstable high-frequency details and inconsistent local appearance across viewpoints. Fresco++ addresses these challenges by regulating both the progression of visual detail and the formation of cross-view supervision during optimization. For frequency-aware optimization, a progressive curriculum first stabilizes low-frequency structures and then introduces high-frequency constraints to recover fine facial and hair details without amplifying spurious responses at early stages. For cross-view optimization, we introduce Canonical Group Consensus, which associates local observations through shared canonical surface regions and establishes correspondence across different viewpoints. Geometric and visibility-aware screening removes unreliable observations, while the remaining multi-view evidence is aggregated in feature space to form a consensus target for supervising the current rendering. This design enforces local consistency without relying on a specific image-space parameterization and avoids additional rendering of the auxiliary view. Together, the frequency curriculum and canonical consensus provide stable optimization from coarse structures to fine details while maintaining coherent appearance across viewpoints. Extensive experiments on NeRSemble demonstrate improved reconstruction quality and cross-view consistency, while evaluations across diverse avatar representations further confirm the generality and transferability of Fresco++.
Read the full abstract

All authors Shikun Zhang, Yong Li, Yiqun Wang, Qiuhong Ke, Cunjian Chen

Computer Vision

GroundSight at GroundLM 2026 Shared Tasks: GoldenViewVQA

Kun Wang, Yupeng Hu, Ruping Cao et al. (+4 more)

GoldenViewVQA requires models to jointly answer driving-scene questions and identify the camera view containing the supporting visual evidence, making precise evidence localization as important as answer correctness. We present CoVeR-VQA, a training-free multi-stage verification and correction framework for grounded multi-view VQA. Starting from GPT-5.6 zero-shot predictions, CoVeR-VQA progressively applies view-specific verification with Gemini-3.6-Flash, prior-guided joint verification with Claude-Opus-5, and cross-split group-level verification that exploits semantically filtered question groups from shared multi-view scenes and validation-derived prior knowledge. On the official GoldenViewVQA test set, the four-stage CoVeR-VQA pipeline achieves 84.75% Joint Accuracy, improving the GPT-5.6 zero-shot baseline by 13.56 percentage points, while reaching 94.92% Answer Accuracy and 86.44% View Accuracy. The final submitted run achieves 88.14% Joint Accuracy after two additional evaluator-informed post-hoc corrections. Our analysis shows that supporting-view localization remains the primary source of residual errors, highlighting the importance of explicit evidence verification for reliable multi-view multimodal reasoning.
Read the full abstract

All authors Kun Wang, Yupeng Hu, Ruping Cao, Hao Liu, Zhiran Li, Qianlong Xiang, Harry Cheng

Computer Vision · Cross-listed from cs.RO

WAM-Cache: Staleness-Bounded KV Reuse for Efficient World Action Models

Kai Ding, Yang He, Ruijie Quan et al. (+1 more)

World Action Models (WAMs) enable generalist robot manipulation by conditioning an action expert on representations from a pretrained video Diffusion Transformer (DiT). In closed-loop control, the video DiT runs at every chunk to encode the current observation into layerwise key-value (KV) pairs that the action expert queries. This prefill dominates the per-chunk computational cost, yet existing training-free accelerations leave it fully dense. We present WAM-Cache, a training-free framework that retains layerwise key-value representations across chunks and recomputes only a sparse refresh set of tokens. Crucially, we find that the intuitive heuristic of refreshing visually drifted tokens plateaus far below the dense baseline, even with an oracle predicting ground-truth KV drift. Downstream action accuracy is instead governed by where the action expert attends, not by what moved. WAM-Cache therefore selects the refresh set by uniting the action expert's cross-attention with visual latent surprise, complemented by a strict age bound that suppresses compounding error. On Fast-WAM, WAM-Cache cuts video DiT prefill FLOPs by 32-42% across RoboTwin 2.0, LIBERO, and real-world experiments, while staying within 0.7-1.8 percentage points of the dense policy in simulation and 2.5 points on a real robot.
Read the full abstract

All authors Kai Ding, Yang He, Ruijie Quan, Yi Yang

Computer Vision · Cross-listed from cs.LG

CoPoE: Multimodal Fusion via Decomposable Disease-Coordinate Product-of-Experts for Missing-Modality Alzheimer's Diagnosis

Chihun An and Ikbeom Jang

Multimodal Alzheimer's disease (AD) diagnosis benefits from integrating heterogeneous clinical, imaging, genomic, and biomarker evidence, but clinical cohorts frequently suffer from irregular modality missingness. Existing fusion methods often synthesize absent inputs, risking the introduction of artificial surrogates, or pool available signals into uninterpretable latent spaces. We present CoPoE (Disease-Coordinate Product-of-Experts), a disease-coordinate framework that maps multimodal evidence into a structured latent space partitioned into four distinct biological and clinical axes: genetic Risk, molecular Pathology, Neurodegeneration, and clinical Stage (R/P/N/S). Each observed modality parameterizes a diagonal Gaussian expert over the full RPNS vector, and a masked Product-of-Experts architecture fuses only the available modalities. Consequently, absent modalities add no factor to the fusion path, allowing the network to preserve a robust, decomposable posterior for any non-empty modality subset without synthetic imputation in the RPNS path. Through extensive missing-modality experiments on the ADNI dataset, CoPoE achieves the best all-modality performance and the highest mean AUROC across all 15 observed-subset evaluations among standardized missing-modality fusion baselines under a shared non-PET ADNI embedding benchmark, while substantially improving raw-probability ECE, Brier score, and NLL. Furthermore, PET-supervised probing shows evidence enrichment within the pathology (P) block under full modalities, with tau-related signal retained even when direct fluid biospecimen inputs are withheld. Our code is available at https://github.com/labhai/CoPoE.
Read the full abstract

Computer Vision

EvoKnow: Continual Knowledge Evolution for AI-Generated Image Detection

Zhiheng Peng, Wenwei Jin, Yangshi Ge et al. (+3 more)

AI-generated image detectors are commonly trained on fixed generator domains and become difficult to maintain as new generative models emerge. Continual adaptation is challenging because replaying historical generated images is costly, whereas updating shared parameters with limited current-domain data can overwrite prior forensic knowledge. We propose EvoKnow, a replay-free framework that formulates continual AI-generated image detection as forensic knowledge evolution. EvoKnow preserves a shared forensic basis learned from base domains, incrementally adds isolated residual experts for complementary generator-relevant evidence, and retrieves expertise through an Analytical Incremental Router (AIR) updated in closed form from current-stage generated images and accumulated sufficient statistics. Experiments demonstrate effective cross-generator generalization, few-shot expansion, and long-horizon continual adaptation. With ten generated images per arriving generator, EvoKnow achieves 96.70% average accuracy on non-base GenImage generators and 94.48% accuracy on Chameleon without target-benchmark adaptation. Under a strict replay-free continual learning protocol, EvoKnow achieves state-of-the-art continual learning performance, attaining 96.32% mean stage-wise accuracy and 4.32% average forgetting.
Read the full abstract

All authors Zhiheng Peng, Wenwei Jin, Yangshi Ge, Siyu Xia, Jiawei Li, Xu Tang

Computer Vision

FastJEV: Understanding Redundancy for Compact JEV Inference

Jie Ma, Jie Gao, Yihang Liu et al. (+5 more)

JEV models make multimodal decisions by directly scoring candidates. Although the common context is encoded once, candidate evaluation can still repeat matching token histories, duplicate inference states, and execute the full backbone. In this paper, we study these sources of redundancy and present FastJEV for compact candidate evaluation. We jointly organize history reuse and state storage, since sharing computation requires preserving states for later branches. We first introduce shared context anchoring to reuse recurrent initial states and omit unused final recurrent caches. We extend this reuse through candidate prefix sharing, retaining the intermediate states needed by subsequent branches. To further reduce the depth of these paths, we apply decision guided pruning based on relative score changes measured on a small unlabeled set. Our method retains full context encoding and all candidates without additional training. We evaluate FastJEV across three OmniJev model sizes on five public benchmarks and reconstructed LIBERO-10 offline questions. At the selected pruning budgets, the complete method reduces candidate depth by 43.75% to 45.83%, while retaining 93.66% to 97.52% of the original task scores on average across the six evaluation sets. Through controlled experiments, we show how candidate overlap and branching structure affect the execution cost of history reuse. In our implementation, candidate prefix sharing can reduce repeated computation while increasing latency. These findings motivate designing sharing granularity and execution schedules together for efficient JEV inference.
Read the full abstract

All authors Jie Ma, Jie Gao, Yihang Liu, Zhike Qiu, Junle Li, Chongyi Zhuang, Jiayi Ji, Xiaoshuai Sun

Computer Vision

CRISP: Fixing Flying Pixels in Latent LiDAR Generation via Diffusion Decoding

Andrea Ceron, Michael Schmidt, Alvaro Marcos-Ramiro et al. (+2 more)

Latent LiDAR pipelines suffer from flying pixels: convolutional VAEs blur sharp radial depth discontinuities, yielding edge depths that back-project to points floating between surfaces. We identify this as a major, directly correctable decoder bottleneck and introduce CRISP: a pixel-space diffusion decoder with a backbone-agnostic latent adapter, DiT-based denoiser, and support mask predictor. CRISP replaces video-VAE and LiDAR-native decoders alike while keeping the encoder and latent generator fixed. Across KITTI-360, SemanticKITTI, and nuScenes, replacing only the decoder reduces FSVD/FPVD by 50.5% on average across frozen backbones; for generic video VAEs, the reductions reach 71%/74%. On the LiDAR-native LiDM backbone, FRID drops by 71%, with the largest gains at depth discontinuities. In a pretrained LiDM world model, the same zero-shot replacement improves FSVD by 15.5%, narrowing the sim-to-real gap.
Read the full abstract

All authors Andrea Ceron, Michael Schmidt, Alvaro Marcos-Ramiro, Sebastian Schmidt, Benjamin Busam

Computer Vision

Rethinking Contrastive Loss in CLIP Post-training: A Complementary Framework with Frozen Text Encoder

Zidan Wang, Yaqian Li, Xiaokai Zhang et al. (+3 more)

CLIP serves as a foundational vision-language model and the de facto vision encoder for downstream VLMs such as LLaVA. Post-training offers a lightweight route to refine CLIP, but recent work argues that the standard contrastive loss is unsuitable for post-training due to catastrophic forgetting under small batches, motivating designs that abandon the contrastive objective in favor of distillation. We revisit this premise and find that, for the InfoNCE objective, the reported forgetting is driven primarily not by insufficient negatives but by an inappropriate magnitude of the contrastive temperature $\tau$: with $\tau$ set sufficiently small, contrastive post-training improves rather than degrades the pretrained CLIP, which we explain through the temperature dependence of the InfoNCE gradient. Building on this finding, we propose ComCLIP, a lightweight single-epoch post-training recipe that freezes CLIP's text encoder—so the refined vision encoder is a drop-in replacement with unchanged architecture and inference cost—and trains the vision encoder with a properly-tempered contrastive loss, an MSE anchoring loss against the original CLIP, and a relational distillation loss from DINOv2. Over multiple seeds, ComCLIP matches the self-distillation baseline CLIP-Refine on zero-shot classification while significantly improving the transferability of visual features, measured by linear probing ($48.99$ vs.\ $42.28$ on ViT-B/16), and on ViT-L/14 it also improves MMVP over CLIP-Refine ($24.20$ vs.\ $19.01$); CLIP-Refine remains stronger on image-text retrieval. Used as a drop-in vision encoder for LLaVA-1.5-7B without re-aligning the projector or LLM, ComCLIP yields no net change across $8$ VLM benchmarks, i.e., the refinement does not break downstream compatibility. Code and models are available at https://github.com/showstarpro/ComCLIP.git.
Read the full abstract

All authors Zidan Wang, Yaqian Li, Xiaokai Zhang, Kaiwen Long, Kun He, Hanpeng Liu

Computer Vision · Cross-listed from cs.CL

SignRAG: Unified Retrieval-Augmented Gloss-Free Sign Language Translation

Zhi Rao, Yucheng Zhou, Qianran Sun et al. (+8 more)

Contemporary decoder-only large language models (LLMs) have demonstrated strong capabilities across a wide range of domains. However, existing pretraining paradigms for gloss-free sign language translation (SLT) are largely designed around conventional encoder-decoder pretrained language models, which limits their direct applicability to decoder-only LLMs. To address this limitation, we propose SignRAG, a unified framework combining hierarchical pretraining, target-domain retrieval augmentation, and retrieval-aware reinforcement fine-tuning. Hierarchical pretraining first learns linguistically grounded sign representations and then jointly aligns the sign encoder with an LLM, mitigating cross-modal optimization imbalance. For downstream adaptation, SignRAG complements parameter-based fine-tuning with a target-domain retrieval gallery that provides instance-specific translation cues. To ensure that retrieved contexts are used appropriately, we further introduce Retrieval Utility-Guided Reinforcement Fine-Tuning (RUG-RFT), which combines translation-quality and retrieval-utility rewards to encourage beneficial retrieval use while suppressing harmful reliance. Experiments on multiple SLT benchmarks establish new state-of-the-art performance. In particular, to the best of our knowledge, SignRAG is the first gloss-free approach to outperform gloss-supervised methods across all reported metrics on CSL-Daily. Our code has been released at \href{https://github.com/shahelaojieraozhi/SignRAG}{GitHub}, together with models of different sizes to support future academic research.
Read the full abstract

All authors Zhi Rao, Yucheng Zhou, Qianran Sun, Yiqing Huang, Longcan Yuan, Jiayi Hou, Chengwen Yao, Lin Cheng, Donghui Sun, Xiaoxin Chen, Jun Wan

Computer Vision

FlyMark: Training-Free Invisible Watermarking of 3D Gaussian Splatting via a Fruit Fly Connectome

Ziyuan Luo, Haoliang Li and Renjie Wan

A trained 3D Gaussian Splatting (3DGS) scene ships as a portable parameter array that can be copied, pruned, requantized, or repackaged outside its training pipeline, so ownership evidence is most useful when it lives in the released parameters and remains checkable long after the embedding tooling is gone. Existing 3DGS watermarks typically tie embedding or extraction to scene optimization, a learned decoder, or rendered views, so the evidence survives only as long as a second trained artifact does. FlyMark instead writes a keyed message into the parameters a 3DGS file already stores. Its carrier directions are derived from the photoreceptors of a published connectome, a citable versioned artifact that fixes the geometry exhaustively and leaves nothing to tune per scene. A virtual observer reads cone-wise apparent luminance along a scene-normalized orbit from stored centers, colors, and opacities; a keyed dithered quantization-index-modulation code replicates each message bit across these observations; and one sparse bounded least-squares solve realizes the targets through achromatic shifts of existing degree-zero colors under a hard per-channel linear-RGB bound. All geometry and higher-order appearance parameters are preserved bit-identically, and extraction needs only cone queries, rounding, and majority voting. Under a model-domain threat model on synthetic and real scenes, FlyMark attains high clean bit accuracy and visual fidelity while cleanly separating matched from wrong keys.
Read the full abstract

Computer Vision · Cross-listed from cs.LG

Bernoulli Flow Models: Self-Consistent Generative Modeling for Binary Data

Hao Mo, Liying Yang, Shumin Yao et al. (+4 more)

Binary diffusion models typically require a large number of function evaluations (NFEs) to generate high-quality samples, making practical inference computationally expensive. Reducing NFEs while preserving sample quality without distillation or additional training remains a significant challenge. Existing binary diffusion models define a discrete one-step forward path and then derive the reverse posterior. In low-NFE settings requiring cross-step sampling, they approximate the true multi-step likelihood with a single-step likelihood transition, which severely degrades sample quality. To address this fundamental limitation and decouple the generative dynamics from fixed discrete time steps, we propose Bernoulli Flow Models (BFM). Rather than relying on sequential one-step Markov diffusion chains, BFM defines a unified continuous global Bernoulli probability flow path between data distributions and pure noise, from which we derive analytical closed-form posterior transitions over arbitrary time intervals. Consequently, reducing the inference NFE is no longer an approximation based on skipping discrete steps; it only requires re-evaluating the analytical posterior over a new time grid. This eliminates the structural training-inference mismatch inherent to discrete chains and yields self-consistent low-NFE sampling. Experiments show that BFM is highly robust to aggressive NFE reduction. On LSUN Churches 256x256, a BFM trained with 256 steps achieves an FID of 9.22 using only 16 sampling steps, whereas the state-of-the-art discrete baseline degrades to 204.10. BFM also remains competitive with continuous and discrete generative baselines under standard full-step inference. These results establish BFM as a theoretically rigorous, self-consistent, and practically effective framework for fast binary data generation.
Read the full abstract

All authors Hao Mo, Liying Yang, Shumin Yao, Xinxing Yu, Ajian Liu, Xudong Mao, Yanyan Liang

Computer Vision · Cross-listed from cs.SD

SepGen: Multi-Stem Audio-Video Separation and Generation in a Single Model

Aviad Dahan, Rajaei Khatib, Yonatan Bitton et al. (+3 more)

A 4D audio-visual scene comprises a video, the dynamic geometry it depicts, and the sound sources that populate it, each with its own position and trajectory. Rendering such a scene from a novel viewpoint requires that every source be available as an individual waveform, so that it can be localized in the scene and propagated to the observer before the signals are mixed. Joint audio-video generators can synthesize the video and its soundtrack, but the soundtrack is emitted as a single audio-mix in which the sources are not individually accessible. We present SepGen, which extends a pretrained audio-video generator to emit the video, the mixed soundtrack, and one waveform per captioned source in one joint sampling run. SepGen supports two complementary modes: generation and separation. In generation mode, each source caption specifies what its stem contains. A two-speaker dialogue, for example, comes out as one stem per speaker in the original turn order. In separation mode, the input audio-mix remains clean while the captions specify what to extract, so the model can decompose a recording from a free-text description. We evaluate generation on scenes synthesized from text, and separation on scenes rendered by other generators and on real recordings of speech, music, and sound effects. Given an audio-mix and captions that carry the spoken lines, SepGen outperforms language-conditioned separators, most clearly on speech, and it keeps the lead when the lines are removed from the captions. Code, checkpoints, and datasets are available at https://sepgen.github.io/
Read the full abstract

All authors Aviad Dahan, Rajaei Khatib, Yonatan Bitton, Idan Szpektor, Lior Wolf, Raja Giryes

Computer Vision

Adaptive Adversarial Augmentation for Controllable Face Synthesis

Saransh Suri, Shivang Agarwal, Mayank Vatsa et al. (+1 more)

Synthetic data provides a scalable alternative to real-world datasets for training face recognition models, particularly under challenging conditions such as low resolution, occlusion, and masks. Yet, most approaches lack diversity and fail to generalize effectively. We propose Ensemble Feedback Controllable Synthesis (EFCS), a guided framework that generates diverse and challenging samples while preserving visual realism. EFCS expands distributional variability, often reflected in higher FID and KID scores compared to single-feedback and random synthesis, while maintaining high precision. Recognition models trained on EFCS data consistently outperform baselines across multiple benchmarks, showing improved generalization to real-world scenarios. Furthermore, we introduce an analytically motivated formulation linking perturbation-induced difficulty, sample utility, and performance degradation, offering principled insights into balancing synthetic data complexity for optimal training. Together, these contributions establish EFCS as an effective and analytically grounded approach for bridging the gap between synthetic and real datasets.
Read the full abstract

All authors Saransh Suri, Shivang Agarwal, Mayank Vatsa, Richa Singh

Computer Vision

EgoPhys: Estimating Peak Contact Force and Mechanical Work from Egocentric Manipulation Video

Zhuo Dong, Jianhua Yang, Haohao Li et al. (+4 more)

Physically grounded manipulation of articulated objects requires understanding both the maximum forces encountered during contact and the work performed as their parts move. Peak contact force and mechanical work quantify these complementary aspects, but estimating them from egocentric video is challenging because physical interaction cues are local and indirect. Moreover, peak force is associated with brief contact events, whereas mechanical work depends on force-motion coupling throughout the contact duration. To address these challenges, we propose EgoPhys, an RGB-only framework comprising Contact-Aware Spatial Aggregation (CASA) and Target-Specific Multi-Expert Temporal Routing (TMTR). CASA integrates appearance and geometry features to emphasize interaction-relevant cues, while TMTR models semantic, event, and motion cues with specialized temporal experts and routes them separately for force and work prediction. On the test split from Hoi! dataset, EgoPhys substantially improves predictions of peak force and mechanical work, achieving MAEs of \(5.205 \pm 0.584\) $N$ and $0.894 \pm 0.081$ $J$, respectively.
Read the full abstract

All authors Zhuo Dong, Jianhua Yang, Haohao Li, Yumeng Zhao, Keji He, Yan Huang, Liang Wang

Computer Vision

Point-Focused Attention Meets Context-Scan State Space: Robust Biological Visual Perception for Point Cloud Representation

Kanglin Qu, Pan Gao, Qun Dai et al. (+1 more)

Synergistically capturing intricate local structures and global contextual dependencies has become a critical challenge in point cloud representation learning. To address this, we introduce PointLearner, a point cloud representation learning network that closely aligns with biological vision which employs an active, foveation-inspired processing strategy, thus enabling local geometric modeling and long-range dependency interactions simultaneously. Specifically, we first design a point-focused attention, which simulates foveal vision at the visual focus through a competitive normalized attention mechanism between local neighbors and spatially downsampled features. The spatially downsampled features are extracted by a pooling method based on learnable inducing points, which can flexibly adapt to the non-uniform distribution of point clouds as the number of inducing points is controlled and they interact directly with point clouds. Second, we propose a context-scan state space that mimics eye's saccade inference, which infers the overall semantic structure and spatial content in the scene through a scan path guided by the Hilbert curve for the bidirectional S6. With this focus-then-context biomimetic design, PointLearner demonstrates remarkable robustness and achieves state-of-the-art performance across multiple point cloud tasks.
Read the full abstract

All authors Kanglin Qu, Pan Gao, Qun Dai, Yuanhao Sun

Computer Vision

PathLang: A Language-Centered Benchmark for Vision-Language Models in Computational Pathology

Fanqi Cheng, Kuo Gong, Shangke Liu et al. (+15 more)

Pathology vision-language models (VLMs) have shown strong visual perception ability, but their robustness in the language domain remains poorly characterized. Existing pathology VLM benchmarks largely rely on canonical closed-set prompts or perturb only generic templates, treating language as a fixed evaluation component rather than a variable axis of model behavior. In clinical practice, however, diagnostic language varies across reports, institutions, and candidate diagnoses. We introduce PathLang, a language-centered and clinically grounded zero-shot benchmark. PathLang holds the underlying slides, ground-truth labels, and image-text evaluation direction fixed while systematically varying only the diagnostic language, so that performance differences reflect how a diagnosis is phrased rather than what is imaged. The language variation follows how pathologists actually rephrase diagnoses (terminology, specificity, and reporting style), and all prompts and candidate pools are validated by six board-certified pathologists. PathLang covers four task families: (1) zero-shot classification with image-text alignment analysis, (2) cross-modal retrieval, (3) paraphrase robustness, including semantic-equivalence paraphrases, length and reporting-style variation, and prompt ensembling, and (4) open-vocabulary diagnosis retrieval over four candidate pools with distinct forms of semantic competition. Across nine VLMs and five public datasets spanning four organs, we find that performance is highly sensitive to clinically equivalent paraphrases, varies substantially across forms of semantic competition, and that image-text alignment quality does not necessarily translate into inter-class separability. We release the prompt corpus, candidate pools, pre-computed text embeddings, and evaluation code at https://anonymous.4open.science/r/PathLang.
Read the full abstract

All authors Fanqi Cheng, Kuo Gong, Shangke Liu, Beidi Zhao, Junchao Zhu, Zheyu Zhu, Leiyue Zhao, Fengbei Liu, John Cannon, Gang Wang, Zu-hua Gao, Kenji Ikemura, Yihe Yang, Yaohong Wang, Yuankai Huo, Xiaoxiao Li, Mert R. Sabuncu, Ruining Deng

Computer Vision

It's Always 10:10: Reference Images Break a Bias That Prompts Only Dent

Luca Cazzaniga

Text-to-image models appear to reproduce the habits of the photographs they learned from. Analog clocks are an extreme case: in advertising, watches almost always show 10:10, and generated clocks return to 10:10 even when another time is requested. We measure this bias and test three ways of overcoming it on 52 models available on the Magnific platform, with a replication on Higgsfield. Every image shows three identical clocks that must show 2:35, 6:50 and 11:20. The description of the object is fixed and only the request about the time changes: no time (A), the time in digits (B), the hand positions described by construction relative to the dial numerals (C), or the same description plus a drawn reference dial (D). Two AI readers read all 1,799 images blind from coded copies, with a third reader and the author settling disagreements (dial-level agreement 96.0% and 97.3%). With no time requested, 67% of the images have all three clocks at 10:10. On the 20 current models, all three clocks are correct in 34% of the images with digits, 30% with the hands described in words and 75% with the reference dial (D-B: +37 points, 95% CI +28 to +45); we found no evidence that describing the hands in words beats the digits (C-B: -4 points, CI -10 to +1). The replication on the 12 models shared by both platforms gives the same picture (B 54%, C 50%, D 81%). Writing the time reduces the bias but leaves two thirds of the images of the 20 current models with at least one wrong clock; adding a drawn reference raises full accuracy to three quarters and almost eliminates images entirely at 10:10. We release all images, prompts, raw readings and a script that recomputes every result.
Read the full abstract

Computer Vision · Cross-listed from cs.LG

Deflating the Hessian: Rank-4 W4A4 Quantization for Multimodal Diffusion Transformers

Shiwen Wang, Pengxiang Zhao and Xiaoming Yuan

In diffusion transformers, low-rank branches can mitigate 4-bit weight–activation (W4A4) post-training quantization (PTQ) loss by decomposing each weight into a low-bit residual and a high-precision low-rank component. Existing low-rank PTQ approaches, however, either optimize low-rank compensation and residual quantization separately, often requiring higher ranks, or rely on second-order weight updates without explicitly modeling activation quantization error, which becomes particularly pronounced under 4-bit quantization. To address these limitations, we present HSVDQuant, a unified framework modeling low-rank-assisted W4A4 PTQ as a coupled calibration problem and deriving optimization-based solvers from the joint objective. Eliminating the output-side low-rank factor yields a deflated Hessian that discounts residual errors already captured by the low-rank component, while an activation-noise surrogate is incorporated to suppress activation quantization error. Across five diffusion backbones, rank-4 HSVDQuant consistently outperforms rank-4 SVDQuant in PSNR and LPIPS. It further surpasses rank-32 SVDQuant on SANA-1.6B, FLUX.1-schnell, and FLUX.1-dev with an $8\times$ smaller rank and up to $6.25\times$ faster quantization. Furthermore, on the Qwen3-8B LLM, rank-4 HSVDQuant improves MMLU accuracy from 61.50% to 68.17% over rank-32 SVDQuant. Overall, HSVDQuant achieves better W4A4 performance with substantially lower rank and quantization cost.
Read the full abstract

Computer Vision

FloorSAV: Elucidating Spatial Audio-Visual Context with 2D Floormap for AV-LLMs

Kyeong-Rae Kim, Sungnyun Kim and Tae-Hyun Oh

While 3D spatial reasoning in dynamic egocentric environments is crucial for embodied intelligence, audio-visual large language models (AV-LLMs) lack explicit mechanisms to process and internalize global geometry directly from raw sensory streams. Existing approaches either require costly fine-tuning or underutilize the model's cross-modal reasoning capacities. In this paper, we propose FloorSAV, a novel framework that explicitly grounds spatial audio-visual context by rendering a dynamic 2D floormap. By integrating 3D point clouds, camera trajectories, spatial audio cues, and semantically grounded object landmarks, we inject this floormap into the AV-LLM as a synchronized stream with an egocentric video. AV-LLMs utilize their multi-modal capabilities to jointly reason over visual, auditory, and geometric cues in a single inference with floormap interpretation guidance. We further introduce SAVED-Bench (Spatial Audio-Visual Egocentric Benchmark with Dynamic Agents), constructing essential tasks of spatial capability in real-world scenarios: dynamic relativity, regional, and path reasoning QAs. FloorSAV improves AV-LLMs' spatial reasoning on various tasks from both SAVED-Bench and SAVVY-Bench. Studies with ground-truth floormaps demonstrate the substantial potential of FloorSAV with accurate spatial information.
Read the full abstract

Computer Vision

Collaboratively Guided Adversarial Robust Distillation with Teacher-Favorable Examples

Zhi Li, Haowei Liu, Hongchen Yang et al. (+4 more)

Adversarial distillation transfers robustness from high-capacity teachers to compact students. Existing adversarial distillation methods mainly use teacher predictions on clean or adversarial examples to supervise student learning. However, teacher-favorable supervision within the perturbation neighborhood remains underexplored in adversarial distillation. We therefore propose Collaboratively Guided Adversarial Robust Distillation (CGARD), which jointly optimizes distinct student-adversarial and teacher-collaborative examples within the same perturbation neighborhood. The teacher-collaborative example is constrained to incur no greater cross-entropy loss under the teacher than the clean input. CGARD combines collaborative teacher guidance with adversarial teacher supervision to improve robust knowledge transfer. Experiments on CIFAR-10 and CIFAR-100, including white-box evaluation and additional black-box transfer evaluation, demonstrate consistent robustness improvements over strong adversarial distillation baselines.
Read the full abstract

All authors Zhi Li, Haowei Liu, Hongchen Yang, Xiaoxuan Wang, Song Gao, Shaowen Yao, Wei Zhou