8829 papers
This paper proposes a two-stage spectral differential privacy framework that fine-tunes a model on low-frequency wavelet coefficients to preserve privacy for sensitive image structures while leveraging public super-resolution models for high-frequency detail, thereby achieving superior image quality compared to standard DP methods.
The paper proposes LineGraph2Road, a novel framework that enhances road network extraction from satellite imagery by transforming keypoint connectivity prediction into a structural reasoning task on line graphs using Graph Transformers, thereby achieving state-of-the-art performance in capturing complex topologies and long-range dependencies.
The paper presents PGVMS, a prompt-guided unified framework that overcomes limitations in virtual multiplex IHC staining by employing adaptive prompt guidance, protein-aware learning, and prototype-consistent learning to generate accurate, spatially aligned multi-stain representations from H&E images using only uniplex training data.
This paper introduces ART-STVG, an auto-regressive transformer architecture with memory selection strategies and a cascaded spatio-temporal design, to effectively address the challenges of localizing targets in long-form videos by processing them as streaming inputs rather than entire sequences.
ManifoldGD is a novel, training-free dataset distillation framework that leverages hierarchical clustering of VAE latent features to construct a multi-scale manifold, guiding diffusion-based synthesis via tangent space projections to generate compact, high-fidelity datasets that outperform existing methods in representativeness and diversity.
PRIMA is a novel multi-modal framework for medical diagnosis that integrates risk-disease correlations via RAG-refined text encoding and a dual-encoder pre-training strategy with specialized loss functions to effectively align visual and clinical metadata, achieving state-of-the-art performance and robustness without requiring massive datasets or extensive computational resources.
This paper proposes a retrieval-augmented test-time adapter that leverages a few-shot support set of pixel-annotated images to fuse textual and visual features, effectively bridging the performance gap between zero-shot and fully supervised open-vocabulary segmentation while preserving the ability to recognize arbitrary categories.
This paper argues that the reasoning limitations of Vision-Language Models stem from a reporting bias in training data that omits tacit information, demonstrating that simply scaling model or data size cannot overcome this issue and that intentional curation of annotations capturing such details is necessary for improvement.
This paper investigates how intrinsic sensor parameters influence event-based object detection models and leverages joint distribution training to achieve sensor-agnostic robustness, thereby addressing data variability and signal characterization gaps in bio-inspired event cameras.
This paper proposes PLADA, a method that drastically reduces dataset transmission costs to under 1 MB by sending only class labels for a preloaded reference dataset, utilizing a pruning mechanism to filter for semantically relevant images and achieve high classification accuracy without transmitting raw pixel data.
The paper proposes SeeThrough3D, a novel model that leverages an occlusion-aware 3D scene representation and masked self-attention to enable text-to-image generation with precise 3D layout control, realistic inter-object occlusions, and consistent camera viewpoints.
VGG-T is a scalable offline feed-forward 3D reconstruction model that overcomes the quadratic computational bottleneck of existing methods by distilling varying-length scene geometry into a fixed-size MLP via test-time training, achieving linear scaling, significantly faster inference, and superior reconstruction accuracy compared to baselines.
MediX-R1 is an open-ended Reinforcement Learning framework that enhances medical multimodal large language models with a composite reward system and LLM-based evaluation to achieve clinically grounded, free-form reasoning, outperforming strong baselines on both text-only and image-text benchmarks despite using a limited dataset.
This paper proposes a novel diffusion model that improves generation speed and image quality by reparameterizing the process using an angular ODE formulation to enable higher-order solvers and by simultaneously estimating both the clean image and noise for more stable inverse diffusion steps.
This paper proposes a real-time motion detection algorithm for streaming video that utilizes Dynamic Mode Decomposition to identify foreground movement by analyzing the correspondence between video feature evolution and the resulting matrix eigenvalues, demonstrating its effectiveness on security footage through ROC analysis and cross-validation.
This paper provides a comprehensive survey on deep learning-based underwater image enhancement by systematically reviewing physical models, data construction, and evaluation metrics; categorizing recent algorithms across six key dimensions; conducting unbiased quantitative and qualitative comparisons of state-of-the-art methods; and outlining future research directions.
This paper introduces the Hallucination benchmark Quality Measurement (HQM) framework to evaluate and improve the reliability and validity of hallucination benchmarks for Large Vision-Language Models, leading to the proposal of the high-quality HQH benchmark which reveals severe hallucination issues in current models and underscores the need for further mitigation.
PoseAdapt is an open-source framework and benchmark suite that enables sustainable human pose estimation by providing standardized protocols for continual learning, allowing models to adapt to changing domains, modalities, and keypoint sets with minimal supervision and without the need for computationally expensive full retraining.
The paper introduces LLM2CLIP, an efficient fine-tuning framework that integrates a large language model into a pretrained CLIP architecture via a lightweight adaptor to significantly enhance cross-modality representation and performance across diverse downstream tasks without requiring large-scale retraining.
This paper introduces Renaissance, a flexible vision-language evaluation framework, to demonstrate through meta-analysis that freezing large parts of models during pretraining significantly reduces compute costs with minimal performance impact and to investigate the effects of initializing vision-language transformers from either vision or text models.