Computer Vision
Papers
Fresh from arXiv, typeset properly. Highlighted words trade on the market.
Friday 14 November
Text-to-image diffusion models generate realistic and coherent images but often fail to follow numerical instructions in text, revealing a gap between language and visual representation. Interestingly, we found that these models are not entirely blind to numbers-they are implicitly aware of their own counting accuracy, as their internal signals shift in consistent ways depending on whether the output meets the specified count. This observation suggests that the model already encodes a latent notion of numerical correctness, which can be harnessed to guide generation more precisely. Building on this intuition, we introduce CountSteer, a training-free method that improves generation of specified object counts by steering the model's cross-attention hidden states during inference. In our experiments, CountSteer improved object-count accuracy by about 4% without compromising visual quality, demonstrating a simple yet effective step toward more controllable and semantically reliable text-to-image generation.
Computer Vision
Toward Gaze Target Detection of Young Autistic Children
The automatic detection of gaze targets in autistic children through artificial intelligence can be impactful, especially for those who lack access to a sufficient number of professionals to improve their quality of life. This paper introduces a new, real-world AI application for gaze target detection in autistic children, which predicts a child's point of gaze from an activity image. This task is foundational for building automated systems that can measure joint attention-a core challenge in Autism Spectrum Disorder (ASD). To facilitate the study of this challenging application, we collected the first-ever Autism Gaze Target (AGT) dataset. We further propose a novel Socially Aware Coarse-to-Fine (SACF) gaze detection framework that explicitly leverages the social context of a scene to overcome the class imbalance common in autism datasets-a consequence of autistic children's tendency to show reduced gaze to faces. It utilizes a two-pathway architecture with expert models specialized in social and non-social gaze, guided by a context-awareness gate module. The results of our comprehensive experiments demonstrate that our framework achieves new state-of-the-art performance for gaze target detection in this population, significantly outperforming existing methods, especially on the critical minority class of face-directed gaze.
Read the full abstract
Computer Vision
Arcee: Differentiable Recurrent State Chain for Generative Vision Modeling with Mamba SSMs
State-space models (SSMs), Mamba in particular, are increasingly adopted for long-context sequence modeling, providing linear-time aggregation via an input-dependent, causal selective-scan operation. Along this line, recent "Mamba-for-vision" variants largely explore multiple scan orders to relax strict causality for non-sequential signals (e.g., images). Rather than preserving cross-block memory, the conventional formulation of the selective-scan operation in Mamba reinitializes each block's state-space dynamics from zero, discarding the terminal state-space representation (SSR) from the previous block. Arcee, a cross-block recurrent state chain, reuses each block's terminal state-space representation as the initial condition for the next block. Handoff across blocks is constructed as a differentiable boundary map whose Jacobian enables end-to-end gradient flow across terminal boundaries. Key to practicality, Arcee is compatible with all prior "vision-mamba" variants, parameter-free, and incurs constant, negligible cost. As a modeling perspective, we view terminal SSR as a mild directional prior induced by a causal pass over the input, rather than an estimator of the non-sequential signal itself. To quantify the impact, for unconditional generation on CelebA-HQ (256$\times$256) with Flow Matching, Arcee reduces FID$\downarrow$ from $82.81$ to $15.33$ ($5.4\times$ lower) on a single scan-order Zigzag Mamba baseline. Efficient CUDA kernels and training code will be released to support rigorous and reproducible research.
Read the full abstract
Computer Vision
Beyond Flatlands: Unlocking Spatial Intelligence by Decoupling 3D Reasoning from Numerical Regression
Existing Vision Language Models (VLMs) architecturally rooted in "flatland" perception, fundamentally struggle to comprehend real-world 3D spatial intelligence. This failure stems from a dual-bottleneck: input-stage conflict between computationally exorbitant geometric-aware encoders and superficial 2D-only features, and output-stage misalignment where discrete tokenizers are structurally incapable of producing precise, continuous numerical values. To break this impasse, we introduce GEODE (Geometric-Output and Decoupled-Input Engine), a novel architecture that resolves this dual-bottleneck by decoupling 3D reasoning from numerical generation. GEODE augments main VLM with two specialized, plug-and-play modules: Decoupled Rationale Module (DRM) that acts as spatial co-processor, aligning explicit 3D data with 2D visual features via cross-attention and distilling spatial Chain-of-Thought (CoT) logic into injectable Rationale Tokens; and Direct Regression Head (DRH), an "Embedding-as-Value" paradigm which routes specialized control tokens to a lightweight MLP for precise, continuous regression of scalars and 3D bounding boxes. The synergy of these modules allows our 1.5B parameter model to function as a high-level semantic dispatcher, achieving state-of-the-art spatial reasoning performance that rivals 7B+ models.
Read the full abstract
Computer Vision
StyleQoRA: Quality-Aware Low-Rank Adaptation for Few-Shot Multi-Style Editing
In recent years, image editing has garnered growing attention. However, general image editing models often fail to produce satisfactory results when confronted with new styles. The challenge lies in how to effectively fine-tune general image editing models to new styles using only a limited amount of paired data and a minimum number of parameters. To address this issue, this paper proposes a novel few-shot multi-style editing framework. For this task, we construct a benchmark dataset that encompasses five distinct styles. Correspondingly, we propose Quality-Aware Low-Rank Adaptation for few-shot multi-style editing (StyleQoRA). Our StyleQoRA can automatically determine the optimal rank for each layer through a novel approach that estimates the importance score of each single-rank component using an image quality metric. To balance specialization and knowledge sharing, we design a Mixture-of-Experts (MoE) LoRA with hybrid routing in our StyleQoRA, consisting of style-specific routing to prevent cross-style confusion and style-shared routing to capture common transformation patterns. Additionally, we explore the optimal location to insert LoRA within the Diffusion in Transformer (DiT) model and integrate adversarial learning and flow matching to guide the diffusion training process. Experimental results demonstrate that our proposed method outperforms existing state-of-the-art approaches with significantly fewer LoRA parameters. Our code and dataset are available at https://github.com/cao-cong/FSMSE.
Read the full abstract
Computer Vision
DoReMi: Bridging 3D Domains via Topology-Aware Domain-Representation Mixture of Experts
Constructing a unified 3D scene understanding model has long been hindered by the significant topological discrepancies across different sensor modalities. While applying the Mixture-of-Experts (MoE) architecture is an effective approach to achieving universal understanding, we observe that existing 3D MoE networks often suffer from semantics-driven routing bias. This makes it challenging to address cross-domain data characterized by "semantic consistency yet topological heterogeneity." To overcome this challenge, we propose DoReMi (Topology-Aware Domain-Representation Mixture of Experts). Specifically, we introduce a self-supervised pre-training branch based on multi attributes, such as topological and texture variations, to anchor cross-domain structural priors. Building upon this, we design a domain-aware expert branch comprising two core mechanisms: Domain Spatial-Guided Routing (DSR), which achieves an acute perception of local topological variations by extracting spatial contexts, and Entropy-controlled Dynamic Allocation (EDA), which dynamically adjusts the number of activated experts by quantifying routing uncertainty to ensure training stability. Through the synergy of these dual branches, DoReMi achieves a deep integration of universal feature extraction and highly adaptive expert allocation. Extensive experiments across various tasks, encompassing both indoor and outdoor scenes, validate the superiority of DoReMi. It achieves 80.1% mIoU on the ScanNet validation set and 77.2% mIoU on S3DIS, comprehensively outperforming existing state-of-the-art methods. The code will be released soon.
Read the full abstract
Computer Vision
DiTGaze: Geometry-Conditioned Diffusion Augmentation for Few-Shot Gaze Estimation
Accurate person-specific gaze estimation benefits from labeled user data, but collecting images that densely cover gaze directions is burdensome. We present DiTGaze, a calibration-efficient augmentation system that follows a two-stream 3D Gaussian Splatting backbone and reformulates its convolutional rendering stage as a geometry-conditioned latent Diffusion Transformer (DiT). A learnable projection and frozen VAE bridge rasterized Gaussian features to latent tokens, while timestep and identity/expression conditioning are integrated through Adaptive Layer Normalization. Intermediate yaw-pitch sampling densifies gaze supervision, and an orthogonality constraint reduces interference between gaze, pose, and expression. On ETH-XGaze, DiTGaze reduces generator-level gaze error by 4.1 percent relative to GazeGaussian. In our runs of the expanded downstream protocol, it ranks first at every calibration budget from one to nine real images; the repeated primary comparison at N = 9 reduces subject-level estimator error by 0.73 degrees and improves 14 of 15 identities (p < 0.001). These results connect the proposed generator changes to statistically supported few-shot personalization gains.
Read the full abstract
Computer Vision
Positional Bias in Multimodal Embedding Models: Do They Favor the Beginning, the Middle, or the End?
Positional bias - where models overemphasize certain positions regardless of content - has been shown to negatively impact model performance across various tasks. While recent research has extensively examined positional bias in text generation models, its presence and effects in representation models remain underexplored. Even less is known about such biases in multimodal models. In this work, we investigate positional bias in multimodal representation models, specifically in the context of image-text retrieval. We begin by distinguishing between context importance and positional bias, and then assess the presence and extent of positional bias across different models and datasets. Our experiments demonstrate that positional bias is prevalent in multimodal models, but manifests differently across modalities: text encoders tend to exhibit bias toward the beginning of the input, whereas image encoders show bias at both the beginning and end. Furthermore, we find that this bias arises from, or is amplified by, a combination of factors, including the positional encoding scheme, training loss, context importance, and the nature of using image-text pairs in multimodal training.
Read the full abstract
Computer Vision
RealisticDreamer: Guidance Score Distillation for Few-shot Gaussian Splatting
3D Gaussian Splatting (3DGS) has recently gained great attention in the 3D scene representation for its high-quality real-time rendering capabilities. However, when the input comprises sparse training views, 3DGS is prone to overfitting, primarily due to the lack of intermediate-view supervision. Inspired by the recent success of Video Diffusion Models (VDM), we propose a framework called Guidance Score Distillation (GSD) to extract the rich multi-view consistency priors from pretrained VDMs. Building on the insights from Score Distillation Sampling (SDS), GSD supervises rendered images from multiple neighboring views, guiding the Gaussian splatting representation towards the generative direction of VDM. However, the generative direction often involves object motion and random camera trajectories, making it challenging for direct supervision in the optimization process. To address this problem, we introduce an unified guidance form to correct the noise prediction result of VDM. Specifically, we incorporate both a depth warp guidance based on real depth maps and a guidance based on semantic image features, ensuring that the score update direction from VDM aligns with the correct camera pose and accurate geometry. Experimental results show that our method outperforms existing approaches across multiple datasets.
Read the full abstract
Computer Vision
MAFM^3: Modular Adaptation of Foundation Models for Multi-Modal Medical AI
Foundational models are trained on extensive datasets to capture the general trends of a domain. However, in medical imaging, the scarcity of data makes pre-training for every domain, modality, or task challenging. Instead of building separate models, we propose MAFM^3 (Modular Adaptation of Foundation Models for Multi-Modal Medical AI), a framework that enables a single foundation model to expand into diverse domains, tasks, and modalities through lightweight modular components. These components serve as specialized skill sets that allow the system to flexibly activate the appropriate capability at the inference time, depending on the input type or clinical objective. Unlike conventional adaptation methods that treat each new task or modality in isolation, MAFM^3 provides a unified and expandable framework for efficient multitask and multimodality adaptation. Empirically, we validate our approach by adapting a chest CT foundation model initially trained for classification into prognosis and segmentation modules. Our results show improved performance on both tasks. Furthermore, by incorporating PET scans, MAFM^3 achieved an improvement in the Dice score 5% compared to the respective baselines. These findings establish that foundation models, when equipped with modular components, are not inherently constrained to their initial training scope but can evolve into multitask, multimodality systems for medical imaging. The code implementation of this work can be found at https://github.com/Areeb2735/CTscan_prognosis_VLM
Read the full abstract
Computer Vision
STONE: Pioneering the One-to-N Universal Backdoor Threat in 3D Point Cloud
Backdoor attacks pose a critical threat to deep learning, especially in safety-sensitive 3D domains such as autonomous driving and robotics. While potent, existing attacks on 3D point clouds are predominantly limited to one-to-one paradigms. The more flexible and universal one-to-N multi-target backdoor threat remains largely unexplored, lacking both theoretical and practical foundations. To bridge this gap, we propose STONE (Spherical Trigger One-to-N universal backdoor Enabling), the first method to instantiate this threat via a configurable spherical trigger design. Its parameterized spatial properties establish a dynamic key space, enabling a single trigger to map to multiple target labels. Theoretically, we ground STONE in a Neural Tangent Kernel (NTK) analysis, providing the first formal basis for one-to-N mappings in 3D models. Empirically, extensive evaluations demonstrate high attack success rates (up to 100%) without compromising clean-data accuracy. This work establishes a foundational benchmark for multi-target backdoor threats under dirty-label and black-box settings in 3D vision – a crucial step toward securing future intelligent systems.
Read the full abstract
Computer Vision
Questioning the Stability of Visual Question Answering
Visual Language Models (VLMs) have achieved remarkable progress, yet their reliability under small, meaning-preserving input changes remains poorly understood. We present the first large-scale, systematic study of VLM robustness to benign visual and textual perturbations: pixel-level shifts, light geometric transformations, padded rescaling, paraphrasing, and multilingual rewrites that do not alter the underlying semantics of an image-question pair. Across a broad set of models and datasets, we find that modern VLMs are highly sensitive to such minor perturbations: a substantial fraction of samples change their predicted answer under at least one visual or textual modification. We characterize how this instability varies across perturbation types, question categories, and models, revealing that even state-of-the-art systems (e.g., GPT-4o, Gemini 2.0 Flash) frequently fail under shifts as small as a few pixels or harmless rephrasings. We further show that sample-level stability serves as a strong indicator of correctness: stable samples are consistently far more likely to be answered correctly. Leveraging this, we demonstrate that the stability patterns of small, accessible open-source models can be used to predict the correctness of much larger closed-source models with high precision. Our findings expose a fundamental fragility in current VLMs and highlight the need for robustness evaluations that go beyond adversarial perturbations, focusing instead on invariances that models should reliably uphold.
Read the full abstract
Computer Vision
Geospatial Chain of Thought Reasoning for Enhanced Visual Question Answering on Satellite Imagery
Geospatial chain of thought (CoT) reasoning is essential for advancing Visual Question Answering (VQA) on satellite imagery, particularly in climate related applications such as disaster monitoring, infrastructure risk assessment, urban resilience planning, and policy support. Existing VQA models enable scalable interpretation of remote sensing data but often lack the structured reasoning required for complex geospatial queries. We propose a VQA framework that integrates CoT reasoning with Direct Preference Optimization (DPO) to improve interpretability, robustness, and accuracy. By generating intermediate rationales, the model better handles tasks involving detection, classification, spatial relations, and comparative analysis, which are critical for reliable decision support in high stakes climate domains. Experiments show that CoT supervision improves accuracy by 34.9% over direct baselines, while DPO yields additional gains in accuracy and reasoning quality. The resulting system advances VQA for multispectral Earth observation by enabling richer geospatial reasoning and more effective climate use cases.
Read the full abstract
Computer Vision
Computationally-efficient deep learning models for nowcasting of precipitation: A solution for the Weather4cast 2025 challenge
This study presents a transfer-learning framework based on Convolutional Gated Recurrent Units (ConvGRU) for short-term rainfall prediction in the Weather4Cast 2025 competition. A single SEVIRI infrared channel (10.8 {\mu}m wavelength) is used as input, which consists of four observations over a one-hour period. A two-stage training strategy is applied to generate rainfall estimates up to four hours ahead. In the first stage, ConvGRU is trained to forecast the brightness temperatures from SEVIRI, enabling the model to capture relevant spatiotemporal patterns. In the second stage, an empirically derived nonlinear transformation maps the predicted fields to OPERA-compatible rainfall rates. For the event-prediction task, the transformed rainfall forecasts are processed using 3D event detection followed by spatiotemporal feature extraction to identify and characterize precipitation events. Our submission achieved 2nd place in the cumulative rainfall task. Further, the same model was used out-of-the-box for the event prediction task, and resulted in similar scores as the baseline model to the competition.
Read the full abstract
Computer Vision
A Comparison of Lightweight Deep Learning Models for Particulate-Matter Nowcasting in the Indian Subcontinent & Surrounding Regions
This paper is a submission for the Weather4Cast~2025 complementary Pollution Task and presents an efficient framework for 6-hour lead-time nowcasting of PM$_1$, PM$_{2.5}$, and PM$_{10}$ across the Indian subcontinent and surrounding regions. The proposed approach leverages analysis fields from the Copernicus Atmosphere Monitoring Service (CAMS) Global Atmospheric Composition Forecasts at 0.4 degree resolution. A 256x256 spatial region, covering 28.4S-73.6N and 32E-134.0E, is used as the model input, while predictions are generated for the central 128x128 area spanning 2.8S-48N and 57.6E-108.4E, ensuring an India-centric forecast domain with sufficient synoptic-scale context. Models are trained on CAMS analyses from 2021-2023 using a shuffled 90/10 split and independently evaluated on 2024 data. Three lightweight parameter-specific architectures are developed to improve accuracy, minimize systematic bias, and enable rapid inference. Evaluation using RMSE, MAE, Bias, and SSIM demonstrates substantial performance gains over the Aurora foundation model, underscoring the effectiveness of compact & specialized deep learning models for short-range forecasts on limited spatial domains.
Read the full abstract
Computer Vision
Viper-F1: Fast and Fine-Grained Multimodal Understanding with Cross-Modal State-Space Modulation
Recent advances in multimodal large language models (MLLMs) have enabled impressive progress in vision-language understanding, yet their high computational cost limits deployment in resource-constrained scenarios such as robotic manipulation, personal assistants, and smart cameras. Most existing methods rely on Transformer-based cross-attention, whose quadratic complexity hinders efficiency. Moreover, small vision-language models often struggle to precisely capture fine-grained, task-relevant visual regions, leading to degraded performance on fine-grained reasoning tasks that limit their effectiveness in the real world. To address these issues, we introduce Viper-F1, a Hybrid State-Space Vision-Language Model that replaces attention with efficient Liquid State-Space Dynamics. To further enhance visual grounding, we propose a Token-Grid Correlation Module, which computes lightweight correlations between text tokens and image patches and modulates the state-space dynamics via FiLM conditioning. This enables the model to selectively emphasize visual regions relevant to the textual prompt while maintaining linear-time inference. Experimental results across multiple benchmarks demonstrate that Viper-F1 achieves accurate, fine-grained understanding with significantly improved efficiency.
Read the full abstract
Computer Vision
Dynamic Gaussian Scene Reconstruction from Unsynchronized Videos
Multi-view video reconstruction plays a vital role in computer vision, enabling applications in film production, virtual reality, and motion analysis. While recent advances such as 4D Gaussian Splatting (4DGS) have demonstrated impressive capabilities in dynamic scene reconstruction, they typically rely on the assumption that input video streams are temporally synchronized. However, in real-world scenarios, this assumption often fails due to factors like camera trigger delays or independent recording setups, leading to temporal misalignment across views and reduced reconstruction quality. To address this challenge, a novel temporal alignment strategy is proposed for high-quality 4DGS reconstruction from unsynchronized multi-view videos. Our method features a coarse-to-fine alignment module that estimates and compensates for each camera's time shift. The method first determines a coarse, frame-level offset and then refines it to achieve sub-frame accuracy. This strategy can be integrated as a readily integrable module into existing 4DGS frameworks, enhancing their robustness when handling asynchronous data. Experiments show that our approach effectively processes temporally misaligned videos and significantly enhances baseline methods.
Read the full abstract
Computer Vision
Refine and Align: Confidence Calibration through Multi-Agent Interaction in VQA
In the context of Visual Question Answering (VQA) and Agentic AI, calibration refers to how closely an AI system's confidence in its answers reflects their actual correctness. This aspect becomes especially important when such systems operate autonomously and must make decisions under visual uncertainty. While modern VQA systems, powered by advanced vision-language models (VLMs), are increasingly used in high-stakes domains like medical diagnostics and autonomous navigation due to their improved accuracy, the reliability of their confidence estimates remains under-examined. Particularly, these systems often produce overconfident responses. To address this, we introduce AlignVQA, a debate-based multi-agent framework, in which diverse specialized VLM – each following distinct prompting strategies – generate candidate answers and then engage in two-stage interaction: generalist agents critique, refine and aggregate these proposals. This debate process yields confidence estimates that more accurately reflect the model's true predictive performance. We find that more calibrated specialized agents produce better aligned confidences. Furthermore, we introduce a novel differentiable calibration-aware loss function called aligncal designed to fine-tune the specialized agents by minimizing an upper bound on the calibration error. This objective explicitly improves the fidelity of each agent's confidence estimates. Empirical results across multiple benchmark VQA datasets substantiate the efficacy of our approach, demonstrating substantial reductions in calibration discrepancies. Furthermore, we propose a novel differentiable calibration-aware loss to fine-tune the specialized agents and improve the quality of their individual confidence estimates based on minimising upper bound calibration error.
Read the full abstract
Computer Vision
CATS-V2V: A Real-World Vehicle-to-Vehicle Cooperative Perception Dataset with Complex Adverse Traffic Scenarios
Vehicle-to-Vehicle (V2V) cooperative perception has great potential to enhance autonomous driving performance by overcoming perception limitations in complex adverse traffic scenarios (CATS). Meanwhile, data serves as the fundamental infrastructure for modern autonomous driving AI. However, due to stringent data collection requirements, existing datasets focus primarily on ordinary traffic scenarios, constraining the benefits of cooperative perception. To address this challenge, we introduce CATS-V2V, the first-of-its-kind real-world dataset for V2V cooperative perception under complex adverse traffic scenarios. The dataset was collected by two hardware time-synchronized vehicles, covering 10 weather and lighting conditions across 10 diverse locations. The 100-clip dataset includes 60K frames of 10 Hz LiDAR point clouds and 1.26M multi-view 30 Hz camera images, along with 750K anonymized yet high-precision RTK-fixed GNSS and IMU records. Correspondingly, we provide time-consistent 3D bounding box annotations for objects, as well as static scenes to construct a 4D BEV representation. On this basis, we propose a target-based temporal alignment method, ensuring that all objects are precisely aligned across all sensor modalities. We hope that CATS-V2V, the largest-scale, most supportive, and highest-quality dataset of its kind to date, will benefit the autonomous driving community in related tasks.
Read the full abstract
Computer Vision
Explainable Deep Convolutional Multi-Type Anomaly Detection
Explainable anomaly detection methods often have the capability to identify and spatially localise anomalies within an image but lack the capability to differentiate the type of anomaly. Furthermore, they often require the costly training and maintenance of separate models for each object category. The lack of specificity is a significant research gap because identifying the type of anomaly (e.g., "Crack" vs. "Scratch") is crucial for accurate diagnosis that facilitates cost-saving operational decisions across diverse application domains. While some recent large-scale Vision-Language Models (VLMs) have begun to address this, they are computationally intensive and memory-heavy, restricting their use in real-time or embedded systems. We propose MultiTypeFCDD, a simple and lightweight convolutional framework designed as a practical alternative for explainable multi-type anomaly detection. MultiTypeFCDD uses only image-level labels to learn and produce multi-channel heatmaps, where each channel is trained to correspond to a specific anomaly type. The model functions as a single, unified framework capable of differentiating anomaly types across multiple object categories, eliminating the need to train and manage separate models for each object category. We evaluated our proposed method on the Real-IAD dataset and it delivers competitive results (96.4% I-AUROC) at just over 1% the size of state-of-the-art VLM models used for similar tasks. This makes it a highly practical and viable solution for real-world applications where computational resources are tightly constrained.
Read the full abstract
Computer Vision
Reverberation: Learning the Latencies Before Forecasting Trajectories
Bridging the past to the future, connecting agents both spatially and temporally, lies at the core of the trajectory prediction task. Despite great efforts, it remains challenging to explicitly learn and predict latencies, i.e., response intervals or temporal delays with which agents respond to various trajectory-changing events and adjust their future paths, whether on their own or interactively. Different agents may exhibit distinct latency preferences for noticing, processing, and reacting to a specific trajectory-changing event. The lack of consideration of such latencies may undermine the temporal continuity of forecasting systems, leading to implausible or unintended trajectories. Inspired by reverberation in acoustics, we propose a new reverberation transform and the corresponding Reverberation (Rev for short) trajectory prediction model, which predicts both individual latency preferences and their stochastic variations accordingly, by using two explicit and learnable reverberation kernels, enabling latency-conditioned and controllable trajectory prediction under both non-interactive and social latencies. Experiments on multiple datasets, whether of pedestrians or vehicles, demonstrate that Rev achieves competitive accuracy while revealing interpretable latency dynamics across agents and scenarios. Qualitative analyses further verify the properties of the reverberation transform, highlighting its potential as a general latency modeling approach.
Read the full abstract
Computer Vision
OT-ALD: Aligning Latent Distributions with Optimal Transport for Accelerated Image-to-Image Translation
The Dual Diffusion Implicit Bridge (DDIB) is an emerging image-to-image (I2I) translation method that preserves cycle consistency while achieving strong flexibility. It links two independently trained diffusion models (DMs) in the source and target domains by first adding noise to a source image to obtain a latent code, then denoising it in the target domain to generate the translated image. However, this method faces two key challenges: (1) low translation efficiency, and (2) translation trajectory deviations caused by mismatched latent distributions. To address these issues, we propose a novel I2I translation framework, OT-ALD, grounded in optimal transport (OT) theory, which retains the strengths of DDIB-based approach. Specifically, we compute an OT map from the latent distribution of the source domain to that of the target domain, and use the mapped distribution as the starting point for the reverse diffusion process in the target domain. Our error analysis confirms that OT-ALD eliminates latent distribution mismatches. Moreover, OT-ALD effectively balances faster image translation with improved image quality. Experiments on four translation tasks across three high-resolution datasets show that OT-ALD improves sampling efficiency by 20.29% and reduces the FID score by 2.6 on average compared to the top-performing baseline models.
Read the full abstract
Computer Vision · Cross-listed from physics.optics
Deep Learning-Enhanced Analysis for Delineating Anticoagulant Essay Efficacy Using Phase Microscopy
The coagulation of blood after it is drawn from the body poses a significant challenge for hematological analysis, potentially leading to inaccurate test results and altered cellular characteristics, compromising diagnostic reliability. This paper presents a deep learning-enhanced framework for delineating anticoagulant efficacy ex vivo using Digital Holographic Microscopy (DHM). We demonstrate a label-free, non-invasive approach for analyzing human blood samples, capable of accurate cell counting and morphological estimation. A DHM with an automated image processing and deep learning pipeline is built for morphological analysis of the blood cells under two different anti-coagulation agents, e.g. conventional EDTA and novel potassium ferric oxalate nanoparticles (KFeOx-NPs). This enables automated high-throughput screening of cells and estimation of blood coagulation rates when samples are treated with different anticoagulants. Results indicated that KFeOx-NPs prevented human blood coagulation without altering the cellular morphology of red blood cells (RBCs), whereas EDTA incubation caused notable changes within 6 hours of incubation. The system allows for quantitative analysis of coagulation dynamics by assessing parameters like cell clustering and morphology over time in these prepared samples, offering insights into the comparative efficacy and effects of anticoagulants outside the body.
Read the full abstract
Computer Vision
From Hindsight to Foresight: Self-Encouraged Hindsight Distillation for Knowledge-based Visual Question Answering
Knowledge-based Visual Question Answering (KBVQA) necessitates external knowledge incorporation beyond cross-modal understanding. Existing KBVQA methods either utilize implicit knowledge in multimodal large language models via in-context learning or explicit knowledge via retrieval augmented generation. However, their reasoning processes remain implicit, without explicit multi-step trajectories. To address this gap, we propose a Self-Encouraged Hindsight Distillation Reasoning (HinD) framework, aiming at eliciting reasoning ability inside the MLLM by constructing a Hindsight Teacher with privileged information to teach the Foresight Student. First, we construct the Hindsight Teacher by prompting the MLLM with the reasoning target as privileged information to complete the reasoning process, obtaining Hindsight-Zero training data. Then, the Foresight Student, without knowing the answer, learns the golden trajectories from Hindsight in two ways: (1) Hindsight Distillation Fine-Tuning to self-distill the Hindsight-Zero into a modularized Chain-of-Thought Generator and a Knowledge Generator for sequential steps and discrete facts generation, respectively; (2) Knowledge Encouragement Preference Optimization to encourage the under-confident but relevant knowledge inside the MLLM and suppress the over-confident but irrelevant one. Experiments on OK-VQA and A-OKVQA validate the effectiveness of HinD, showing that HinD with 7-8B MLLM achieves superior performance without commercial model APIs or retrieved knowledge.
Read the full abstract
Computer Vision · Cross-listed from cs.CL
Enhancing Meme Emotion Understanding with Multi-Level Modality Enhancement and Dual-Stage Modal Fusion
With the rapid rise of social media and Internet culture, memes have become a popular medium for expressing emotional tendencies. This has sparked growing interest in Meme Emotion Understanding (MEU), which aims to classify the emotional intent behind memes by leveraging their multimodal contents. While existing efforts have achieved promising results, two major challenges remain: (1) a lack of fine-grained multimodal fusion strategies, and (2) insufficient mining of memes' implicit meanings and background knowledge. To address these challenges, we propose MemoDetector, a novel framework for advancing MEU. First, we introduce a four-step textual enhancement module that utilizes the rich knowledge and reasoning capabilities of Multimodal Large Language Models (MLLMs) to progressively infer and extract implicit and contextual insights from memes. These enhanced texts significantly enrich the original meme contents and provide valuable guidance for downstream classification. Next, we design a dual-stage modal fusion strategy: the first stage performs shallow fusion on raw meme image and text, while the second stage deeply integrates the enhanced visual and textual features. This hierarchical fusion enables the model to better capture nuanced cross-modal emotional cues. Experiments on two datasets, MET-MEME and MOOD, demonstrate that our method consistently outperforms state-of-the-art baselines. Specifically, MemoDetector improves F1 scores by 4.3% on MET-MEME and 3.4% on MOOD. Further ablation studies and in-depth analyses validate the effectiveness and robustness of our approach, highlighting its strong potential for advancing MEU. Our code is available at https://github.com/singing-cat/MemoDetector.
Read the full abstract
Computer Vision · Cross-listed from cs.CL
AV-Dialog: Spoken Dialogue Models with Audio-Visual Input
Dialogue models falter in noisy, multi-speaker environments, often producing irrelevant responses and awkward turn-taking. We present AV-Dialog, the first multimodal dialog framework that uses both audio and visual cues to track the target speaker, predict turn-taking, and generate coherent responses. By combining acoustic tokenization with multi-task, multi-stage training on monadic, synthetic, and real audio-visual dialogue datasets, AV-Dialog achieves robust streaming transcription, semantically grounded turn-boundary detection and accurate responses, resulting in a natural conversational flow. Experiments show that AV-Dialog outperforms audio-only models under interference, reducing transcription errors, improving turn-taking prediction, and enhancing human-rated dialogue quality. These results highlight the power of seeing as well as hearing for speaker-aware interaction, paving the way for {spoken} dialogue agents that perform {robustly} in real-world, noisy environments.
Read the full abstract
Computer Vision
Stroke Modeling Enables Vectorized Character Generation with Large Vectorized Glyph Model
Vectorized glyphs are widely used in poster design, network animation, art display, and various other fields due to their scalability and flexibility. In typography, they are often seen as special sequences composed of ordered strokes. This concept extends to the token sequence prediction abilities of large language models (LLMs), enabling vectorized character generation through stroke modeling. In this paper, we propose a novel Large Vectorized Glyph Model (LVGM) designed to generate vectorized Chinese glyphs by predicting the next stroke. Initially, we encode strokes into discrete latent variables called stroke embeddings. Subsequently, we train our LVGM via fine-tuning DeepSeek LLM by predicting the next stroke embedding. With limited strokes given, it can generate complete characters, semantically elegant words, and even unseen verses in vectorized form. Moreover, we release a new large-scale Chinese SVG dataset containing 907,267 samples based on strokes for dynamically vectorized glyph generation. Experimental results show that our model has scaling behaviors on data scales. Our generated vectorized glyphs have been validated by experts and relevant individuals.
Read the full abstract
Computer Vision
Toward Generalized Detection of Synthetic Media: Limitations, Challenges, and the Path to Multimodal Solutions
Artificial intelligence (AI) in media has advanced rapidly over the last decade. The introduction of Generative Adversarial Networks (GANs) improved the quality of photorealistic image generation. Diffusion models later brought a new era of generative media. These advances made it difficult to separate real and synthetic content. The rise of deepfakes demonstrated how these tools could be misused to spread misinformation, political conspiracies, privacy violations, and fraud. For this reason, many detection models have been developed. They often use deep learning methods such as Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs). These models search for visual, spatial, or temporal anomalies. However, such approaches often fail to generalize across unseen data and struggle with content from different models. In addition, existing approaches are ineffective in multimodal data and highly modified content. This study reviews twenty-four recent works on AI-generated media detection. Each study was examined individually to identify its contributions and weaknesses, respectively. The review then summarizes the common limitations and key challenges faced by current approaches. Based on this analysis, a research direction is suggested with a focus on multimodal deep learning models. Such models have the potential to provide more robust and generalized detection. It offers future researchers a clear starting point for building stronger defenses against harmful synthetic media.
Read the full abstract
Computer Vision
VIDEOP2R: Video Understanding from Perception to Reasoning
Reinforcement fine-tuning (RFT), a two-stage framework consisting of supervised fine-tuning (SFT) and reinforcement learning (RL) has shown promising results on improving reasoning ability of large language models (LLMs). Yet extending RFT to large video language models (LVLMs) remains challenging. We propose VideoP2R, a novel process-aware video RFT framework that enhances video reasoning by modeling perception and reasoning as distinct processes. In the SFT stage, we develop a three-step pipeline to generate VideoP2R-CoT-162K, a high-quality, process-aware chain-of-thought (CoT) dataset for perception and reasoning. In the RL stage, we introduce a novel process-aware group relative policy optimization (PA-GRPO) algorithm that supplies separate rewards for perception and reasoning. Extensive experiments show that VideoP2R achieves state-of-the-art (SotA) performance on six out of seven video reasoning and understanding benchmarks. Ablation studies further confirm the effectiveness of our process-aware modeling and PA-GRPO and demonstrate that model's perception output is information-sufficient for downstream reasoning. Our project page is available at https://videop2r.github.io/videop2r/.
Read the full abstract
Computer Vision · Cross-listed from cs.MM
AccKV: Towards Efficient Audio-Video LLMs Inference via Adaptive-Focusing and Cross-Calibration KV Cache Optimization
Recent advancements in Audio-Video Large Language Models (AV-LLMs) have enhanced their capabilities in tasks like audio-visual question answering and multimodal dialog systems. Video and audio introduce an extended temporal dimension, resulting in a larger key-value (KV) cache compared to static image embedding. A naive optimization strategy is to selectively focus on and retain KV caches of audio or video based on task. However, in the experiment, we observed that the attention of AV-LLMs to various modalities in the high layers is not strictly dependent on the task. In higher layers, the attention of AV-LLMs shifts more towards the video modality. In addition, we also found that directly integrating temporal KV of audio and spatial-temporal KV of video may lead to information confusion and significant performance degradation of AV-LLMs. If audio and video are processed indiscriminately, it may also lead to excessive compression or reservation of a certain modality, thereby disrupting the alignment between modalities. To address these challenges, we propose AccKV, an Adaptive-Focusing and Cross-Calibration KV cache optimization framework designed specifically for efficient AV-LLMs inference. Our method is based on layer adaptive focusing technology, selectively focusing on key modalities according to the characteristics of different layers, and enhances the recognition of heavy hitter tokens through attention redistribution. In addition, we propose a Cross-Calibration technique that first integrates inefficient KV caches within the audio and video modalities, and then aligns low-priority modalities with high-priority modalities to selectively evict KV cache of low-priority modalities. The experimental results show that AccKV can significantly improve the computational efficiency of AV-LLMs while maintaining accuracy.