Single-view face relighting aims to adjust the portrait lighting while preserving the original background. Although recent diffusion-based methods achieve great relit results by using reference lighting and facial features as conditions for the diffusion relighting process, they are limited by the incompleteness of these conditions, such as the absence of explicit constraints on skin tone and the lack of spatial coverage for hard shadows, which results in facial and lighting inconsistencies. To address these challenges, we propose IFS-Light, an interactive framework that leverages spatial-nonspatial conditioning mechanism to localize facial features and reference lighting, then optimize their interplay in the relighting process. To ensure facial consistency, we first combine skin-tone-scaled conditions with shape information for tone adjustment, enhanced by a detail mask that identifies modifiable facial regions. Skin and shape parameters are then optimized to preserve both skin tone and fine details. To maintain lighting consistency, we propose a ray-tracing-based formulation that decomposes reference lighting into diffuse and non-diffuse components. These, integrated with shape information, assist in positioning light and shadow regions. Both components are then encoded for precise control over color and intensity. In addition, we propose an innovative and user-friendly solution for adjusting light conditions, which enables the user to precisely adjust the position of the light source to flexibly control both light intensity and direction, thereby making it easier to achieve the desired relighting results. Extensive experiments show that IFS-Light achieves superior relighting results compared to state-of-the-art methods. The code and appendix are available https://github.com/mRobotit/IFS-Light
论文检索
输入标题、作者或关键词，从 1,620 篇学术成果中精准定位
Recent advances in 4D Gaussian Splatting have boosted dynamic scene reconstruction and real-time rendering. However, current methods remain retrospective, lacking the ability to forecast future states-limiting their utility in tasks like autonomous navigation and robotics. To address these limitations, we propose FutureGS, a novel Gaussian-based dynamic scene representation framework tailored for continuous 3D future scene prediction and view synthesis. FutureGS introduces a dual-domain decoupled representation, consisting of a static 3D Gaussian base to maintain spatial consistency and a dynamic deformation field to explicitly model temporal motion evolution. To capture long-range dependencies and complex motion dynamics, we design a multi-window collaborative prediction strategy that leverages a sliding temporal window and a bidirectional LSTM-based temporal encoder for robust future motion estimation. Furthermore, we propose a KNN-based local rigidity-aware fusion mechanism, which adaptively regulates the prediction consistency based on local deformation intensity, enhancing the geometric stability and physical plausibility of future scenes. Extensive experiments on standard dynamic scene benchmarks, including D-NeRF and NeRF-DS, demonstrate that FutureGS achieves superior performance in terms of visual fidelity and spatiotemporal consistency, enabling real-time and photorealistic rendering from arbitrary viewpoints at future time steps.
Cued Speech (CS) is a visual communication system that combines lip-reading with hand coding to facilitate communication for individuals with hearing impairments. Automatic CS Recognition (ACSR) aims to convert CS hand gestures and lip movements into text via AI-driven methods. Traditionally, the temporal asynchrony between hand and lip movements requires the design of complex modules to facilitate effective multimodal fusion. However, constrained by limited data availability, current methods demonstrate insufficient capacity for adequately training these fusion mechanisms, resulting in suboptimal performance. Recently, multi-agent systems have shown promising capabilities in handling complex tasks with limited data availability. To this end, we propose the first collaborative multi-agent system for ACSR, named Cued-Agent. It integrates four specialized sub-agents: a Multimodal Large Language Model-based Hand Recognition agent that employs keyframe screening and CS expert prompt strategies to decode hand movements, a pretrained Transformer-based Lip Recognition agent that extracts lip features from the input video, a Hand Prompt Decoding agent that dynamically integrates hand prompts with lip features during inference in a training-free manner, and a Self-Correction Phoneme-to-Word agent that enables post-processing and end-to-end conversion from phoneme sequences to natural language sentences for the first time through semantic refinement. To support this study, we expand the existing Mandarin CS dataset by collecting data from eight hearing-impaired cuers, establishing a mixed dataset of fourteen subjects. Extensive experiments demonstrate that our Cued-Agent performs superbly in both normal and hearing-impaired scenarios compared with state-of-the-art methods. The implementation is available at https://github.com/DennisHgj/Cued-Agent.
Few-Shot Class-Incremental Learning (FSCIL) aims to continuously learn novel concepts from limited training samples without forgetting previously encountered classes. Recent advancements have leveraged Parameter-Efficient Tuning (PET) strategies on pre-trained models to enhance FSCIL performance. However, current PET-based FSCIL approaches still suffer from the challenges posed by catastrophic collapse of general prompt and limited adaptability of specific prompt . To this end, we redefine the function of the PET paradigm with both gradient-aware prompting (GAP) and router-free adapters (RFA) to boost the performance of FSCIL, termed as "PET-GPRA". To dynamically balance the retention of previously learned general knowledge and the acquisition of novel class information across sessions, the GAP paradigm adaptively adjusts the updated gradient of the general prompt by leveraging the angular relationship between the general knowledge gradient and the novel knowledge gradient. Meanwhile, the RFA mechanism utilizes the semantic similarity between class attributes to replace the routing network, guiding the integration of adapter information, in which adapters serve as specific prompts to enhance the adaptability. Extensive experiments on multiple benchmark datasets consistently demonstrate the superiority and effectiveness of our proposed PET-GPRA framework over state-of-the-art baselines.
Audio phase retrieval aims to reconstruct phase from the given magnitude and obtain the time-domain audio waveform. While deep learning techniques have promoted the development of this area, existing deep neural network (DNN)-based methods usually suffer from some inherent problems like limited generalization capability to different audio types, failing to adapt to different sampling rates, and inflexibility for varying computational complexity during the inference stage, which heavily hinder the development of the filed. To tackle these challenges, in this paper, we introduce a novel phase task estimation task called versatile auido phase retrieval and a Band-Aware Phase Estimation Network (BAPEN ) is proposed. Specifically, we first collect and establish a new benchmark for the task, which encompasses speech, sound effects, and music and the total duration is around 414 hours. Besides, a sub-band oriented framework is proposed, which involves hierarchical sub-band encoding/decoding and a dual-path network structure is specially devised for efficient narrow- and cross-band modeling, respectively. Furthermore, to enable dynamic control over the inference cost, we propose a simple yet effective sampling strategy for network depth augmentation during training. Both objective and subject results validate the promising performance of the BAPEN while possessing more flexible application ranges. Audio samples are available on: https://lingling-dai.github.io/BAPEN/.
DINOv2 Driven Gait Representation Learning for Video-Based Visible-Infrared Person Re-identification
Video-based Visible-Infrared person re-identification (VVI-ReID) aims to retrieve the same pedestrian across visible and infrared modalities from video sequences. Existing methods tend to exploit modality-invariant visual features but largely overlook gait features, which are not only modality-invariant but also rich in temporal dynamics, thus limiting their ability to model the spatiotemporal consistency essential for cross-modal video matching. To address these challenges, we propose a DINOv2-Driven Gait Representation Learning (DinoGRL) framework that leverages the rich visual priors of DINOv2 to learn gait features complementary to appearance cues, facilitating robust sequence-level representations for cross-modal retrieval. Specifically, we introduce a Semantic-Aware Silhouette and Gait Learning (SASGL) model, which generates and enhances silhouette representations with general-purpose semantic priors from DINOv2 and jointly optimizes them with the ReID objective to achieve semantically enriched and task-adaptive gait feature learning. Furthermore, we develop a Progressive Bidirectional Multi-Granularity Enhancement (PBMGE) module, which progressively refines feature representations by enabling bidirectional interactions between gait and appearance streams across multiple spatial granularities, fully leveraging their complementarity to enhance global representations with rich local details and produce highly discriminative features. Extensive experiments on HITSZ-VCM and BUPT datasets demonstrate the superiority of our approach, significantly outperforming existing state-of-the-art methods.
Rheumatoid arthritis (RA) is a chronic autoimmune disease characterized by joint inflammation and progressive structural damage. Joint space width (JSW) is a critical indicator in conventional radiography (CR) for evaluating disease progression, which has become a prominent research topic in computer-aided diagnostic (CAD) systems. However, deep learning-based radiological CAD systems for JSW analysis face significant challenges in data quality, including data imbalance, limited variety, and annotation difficulties. This work introduced a challenging image synthesis scenario and proposed Layer Separation Networks (LSN) to accurately separate the soft tissue layer, the upper bone layer, and the lower bone layer in conventional radiographs of finger joints. Using these layers, the adjustable JSW images can be synthesized to address data quality challenges and achieve ground truth (GT) generation. Experimental results demonstrated that LSN-based synthetic images closely resemble real radiographs, and significantly enhanced the performance in downstream tasks. The code and dataset are available at: https://github.com/pokeblow/LSN.
Event cameras, with their microsecond-level temporal resolution and sparse visual encoding, provide a transformative paradigm for automatic lip reading (ALR). However, event data inherently lack explicit spatial structure and exhibit a pronounced frequency-domain bias. The low-frequency components fail to capture crucial lip structural information, which fundamentally impedes the modeling of intra-frame topological dependencies and inter-frame semantic evolution-both of which are critical for robust lip reading. To this end, we propose FAST-HG, a Frequency-Aware SpatioTemporal HyperGraph framework specifically designed for event-based lip reading. First, we apply low-frequency perturbation to improve the model's robustness for capturing discriminative features, and integrate adaptive high-frequency filtering to enhance edge-aware representations. Then, we construct a Spatial Region Hypergraph (SRH) and a Temporal Semantic Hypergraph (TSH). The former captures intra-frame topological dependencies among lip regions, while the latter explicitly models inter-frame structural associations throughout the lip movement process, enabling the model to capture discriminative patterns in lip dynamics. Furthermore, we propose a viseme-aware label smoothing strategy, where a novel viseme-level edit distance is designed to quantify visual similarities between classes and guide the construction of soft labels. FAST-HG achieves 79.85% and 84.03% accuracy on the DVS-Lip and DVS-LRW100 datasets, respectively, significantly outperforming prior methods and establishing a new benchmark for event-based lip reading.
In filmmaking, directors typically allow actors to perform freely based on the script before providing specific guidance on how to present key actions. AI-generated content faces similar requirements, where users not only need automatic generation of lip synchronization and basic gestures from audio input but also desire semantically accurate and expressive body movement that can be ''directly guided'' through text descriptions. Therefore, we present VersaAnimator, a versatile framework that synthesizes expressive talking human videos from arbitrary portrait images. Specifically, we design a motion generator that produces basic rhythmic movements from audio input and supports text-prompt control for specific actions. The generated whole-body 3D motion tokens can animate portraits of various scales, producing talking heads, half-body gestures and even leg movements for whole-body images. Besides, we introduce a multi-modal controlled video diffusion that generates photorealistic videos, where speech signals govern lip synchronization, facial expressions, and head motions while body movements are guided by the 2D poses. Furthermore, we introduce a token2pose translator to smoothly map 3D motion tokens to 2D pose sequences. This design mitigates the stiffness resulting from direct 3D to 2D conversion and enhances the details of the generated body movements. Extensive experiments shows that VersaAnimator synthesizes lip-synced and identity-preserving videos while generating expressive and semantically meaningful whole-body motions. https://digital-avatar.github.io/ai/VersaAnimator/
Recent advancements in image editing have utilized large-scale multimodal models to enable intuitive, natural instruction-driven interactions. However, conventional methods still face significant challenges, particularly in spatial reasoning, precise region segmentation, and maintaining semantic consistency, especially in complex scenes.To overcome these challenges, we introduce SmartFreeEdit, a novel end-to-end framework that integrates a multimodal large language model (MLLM) with a hypergraph-enhanced inpainting architecture, enabling precise, mask-free image editing guided exclusively by natural language instructions. The key innovations of SmartFreeEdit include: (1) the introduction of region-aware tokens and a mask embedding paradigm that enhance the model's spatial understanding of complex scenes; (2) a reasoning segmentation pipeline designed to optimize the generation of editing masks based on natural language instructions; and (3) a hypergraph-augmented inpainting module that ensures the preservation of both structural integrity and semantic coherence during complex edits, overcoming the limitations of local-based image generation. Extensive experiments on the Reason-Edit benchmark demonstrate that SmartFreeEdit surpasses current state-of-the-art methods across multiple evaluation metrics, including segmentation accuracy, instruction adherence, and visual quality preservation, while addressing the issue of local information focus and improving global consistency in the edited image. Our project will be available at https://github.com/smileformylove/SmartFreeEdit.
Accurate color alignment in text-to-image (T2I) generation is critical for applications such as fashion, product visualization, and interior design, yet current diffusion models struggle with nuanced and compound color terms (e.g., Tiffany blue, baby pink), often producing images that are misaligned with human intent. Existing approaches rely on cross-attention manipulation, reference images, or fine-tuning but fail to systematically resolve ambiguous color descriptions. To precisely render colors under prompt ambiguity, we propose a training-free framework that enhances color fidelity by leveraging a large language model (LLM) to disambiguate color-related prompts and guiding color blending operations directly in the text embedding space. Our method first employs a large language model (LLM) to resolve ambiguous color terms in the text prompt, and then refines the text embeddings based on the spatial relationships of the resulting color terms in the CIELab color space. Unlike prior methods, our approach improves color accuracy without requiring additional training or external reference images. Experimental results demonstrate that our framework improves color alignment without compromising image quality, bridging the gap between text semantics and visual generation. All supplementary materials are available at https://Sung-Lin.github.io/TintBench/.
Remote Sensing Image-Text Retrieval (RSITR) is a fundamental task in the remote sensing (RS) field and has seen significant progress in recent years. However, existing methods often overlook explicit attention to semantic entities in RS scenes, limiting their capabilities in fine-grained semantic modeling and cross-modal matching, thereby hindering retrieval performance. To address these limitations, we propose a novel framework, Entity-level Alignment with Prompt-guided Adapter (EAPA), which enhances retrieval performance by explicitly perceiving, embedding, and aligning semantic entities in RS images and texts. Built upon the Contrastive Language-Image Pretraining (CLIP) model, EAPA comprises three key modules: the Prompt-guided Attention Adapter (PAA) module, the Pseudo-label-supervised Entity Embedding (PEE) module, and the Cross-modal Entity-level Semantic Alignment (CESA) module. Specifically, PAA freezes the CLIP backbone and introduces learnable prompt vectors to capture RS-specific entity-level semantic knowledge, guiding attention distribution and enhancing semantic representations. To obtain cross-modal consistent entity-level representations, PEE employs an entity query-based encoder to extract entity embeddings of both images and texts, and uses pseudo semantic labels as supervision to ensure that each embedding corresponds to a unique and well-defined semantic category. Based on this, CESA performs one-to-one alignment of cross-modal entity embeddings that correspond to the same semantic category, effectively avoiding mismatches and enhancing fine-grained alignment. Extensive experiments on the RSICD and RSITMD datasets demonstrate that EAPA outperforms state-of-the-art methods across multiple metrics, validating the effectiveness of each module in enhancing fine-grained semantic modeling and cross-modal matching.
Recent advances in 3D Gaussian Splatting (3DGS) have greatly improved 3D reconstruction. However, its substantial data size poses a significant challenge for transmission and storage. While many compression techniques have been proposed, they fail to efficiently adapt to fluctuating network bandwidth, leading to resource wastage. We address this issue from the perspective of size-aware compression, where we aim to compress 3DGS to a desired size by quickly searching for suitable hyperparameters. Through a measurement study, we identify key hyperparameters that affect the size - namely, the reserve ratio of Gaussians and bit-width settings for Gaussian attributes. Then, we formulate this hyperparameter optimization problem as a mixed-integer nonlinear programming (MINLP) problem, with the goal of maximizing visual quality while respecting the size budget constraint. To solve the MINLP, we decouple this problem into two parts: discretely sampling the reserve ratio and determining the bit-width settings using integer linear programming (ILP). To solve the ILP more quickly and accurately, we design a quality loss estimator and a calibrated size estimator, as well as implement a CUDA kernel. Extensive experiments on multiple 3DGS variants demonstrate that our method achieves state-of-the-art performance in post-training compression. Furthermore, our method can achieve comparable quality to leading training-required methods after fine-tuning.
Removing various degradations from damaged documents greatly benefits digitization, downstream document analysis, and readability. Previous methods often treat each restoration task independently with dedicated models, leading to a cumbersome and highly complex document processing system. Although recent studies attempt to unify multiple tasks, they often suffer from limited scalability due to handcrafted prompts and heavy preprocessing, and fail to fully exploit inter-task synergy within a shared architecture. To address the aforementioned challenges, we propose Uni-DocDiff, a Unified and highly scalable Doc ument restoration model based on Dif fusion. Uni-DocDiff develops a learnable task prompt design, ensuring exceptional scalability across diverse tasks. To further enhance its multi-task capabilities and address potential task interference, we devise a novel Prior Pool, a simple yet comprehensive mechanism that combines both local high-frequency features and global low-frequency features. Additionally, we design the Prior Fusion Module (PFM), which enables the model to adaptively select the most relevant prior information for each specific task. Extensive experiments show that the versatile Uni-DocDiff achieves performance comparable or even superior performance compared with task-specific expert models, and simultaneously holds the task scalability for seamless adaptation to new tasks.
Semi-supervised Camouflaged Object Detection (SSCOD) aims to reduce reliance on costly pixel-level annotations by leveraging limited annotated data and abundant unlabeled data. However, existing SSCOD methods based on Teacher-Student frameworks suffer from severe prediction bias and error propagation under scarce supervision, while their multi-network architectures incur high computational overhead and limited scalability. To overcome these limitations, we propose ST-SAM, a highly annotation-efficient yet concise framework that breaks away from conventional SSCOD constraints. Specifically, ST-SAM employs Self-Training strategy that dynamically filters and expands high-confidence pseudo-labels to enhance a single-model architecture, thereby fundamentally circumventing inter-model prediction bias. Furthermore, by transforming pseudo-labels into hybrid prompts containing domain-specific knowledge, ST-SAM effectively harnesses the Segment Anything Model's potential for specialized tasks to mitigate error accumulation in self-training. Experiments on COD benchmark datasets demonstrate that ST-SAM achieves state-of-the-art performance with only 1% labeled data, outperforming existing SSCOD methods and even matching fully supervised methods. Remarkably, ST-SAM requires training only a single network, without relying on specific models or loss functions. This work establishes a new paradigm for annotation-efficient SSCOD. Codes will be available at https://github.com/hu-xh/ST-SAM.
Image degradation caused by complex lighting conditions such as low-light and backlit scenarios is commonly encountered in real-world environments, significantly affecting image quality and downstream vision tasks. Most existing methods focus on a single type of illumination degradation and lack the ability to handle diverse lighting conditions in a unified manner. To address this issue, we propose a dual-illumination enhancement framework called DIME-Net. The core of our method is a Mixture-of-Experts illumination estimator module, where a sparse gating mechanism adaptively selects suitable S-curve expert networks based on the illumination characteristics of the input image. By integrating Retinex theory, this module effectively performs enhancement tailored to both low-light and backlit images. To further correct illumination-induced artifacts and color distortions, we design a damage restoration module equipped with Illumination-Aware Cross Attention and Sequential-State Global Attention mechanisms. In addition, we construct a hybrid illumination dataset, MixBL, by integrating existing datasets, allowing our model to achieve robust illumination adaptability through a single training process. Experimental results show that DIME-Net achieves competitive performance on both synthetic and real-world low-light and backlit datasets without any retraining. These results demonstrate its generalization ability and potential for practical multimedia applications under diverse and complex illumination conditions.
Visual selective attention, driven by individual preferences, regulates human prioritization of visual stimuli by bridging subjective cognitive mechanisms with objective visual elements, thereby steering the semantic interpretation and hierarchical processing of dynamic visual scenes. However, existing models and datasets predominantly neglect the influence of subjective cognitive diversity on fixation behavior. Conventional saliency prediction models, typically employing segmentation approaches, rely on low-resolution imagery to generate saliency heatmaps, subsequently upscaled to native resolutions, which limiting their capacity to capture personalized attention patterns. Furthermore, MLLMs are constrained by factors such as hallucinations, making it very costly to strictly adhere to the expected format in tasks involving multiple point predictions, and achieving precise point positioning is challenging. To address these limitations, we present Subjective Personalized Attention for Ad vertisement Videos, namely SPA-ADV, a large-scale multimodal dataset capturing gaze behaviors from over 4,500 participants varying in age and gender with 486 videos. Furthermore, we propose PRE-MAP, a novel eye-tracking saliency model that characterizes Personalized visual disparities through Reinforcement learning-optimized Eye-tracking, built upon MLLMs and guided by Multi-Attribute user profiles to predict Points. To ensure MLLMs produce prediction points that are both format-correct and spatially accurate, we introduce Consistency Group Relative Policy Optimization (C-GRPO), inspired by the variability in eye movement points and Multi-Attribute profiles. Extensive experiments on SPA-ADV and other benchmarks demonstrate the effectiveness of our approach. The code and dataset are available at https://github.com/mininglamp-MLLM/PRE-MAP.
In current web environment, fake news spreads rapidly across online social networks, posing serious threats to society. Existing multimodal fake news detection methods can generally be classified into knowledge-based and semantic-based approaches. However, these methods are heavily rely on human expertise and feedback, lacking flexibility. To address this challenge, we propose a Dynamic Analysis and Adaptive Discriminator (DAAD) approach for fake news detection. For knowledge-based methods, we introduce the Monte Carlo Tree Search algorithm to leverage the self-reflective capabilities of large language models (LLMs) for prompt optimization, providing richer, domain-specific details and guidance to the LLMs, while enabling more flexible integration of LLM comment on news content. For semantic-based methods, we define four typical deceit patterns: emotional exaggeration, logical inconsistency, image manipulation, and semantic inconsistency, to reveal the mechanisms behind fake news creation. To detect these patterns, we carefully design four discriminators and expand them in depth and breadth, using the soft-routing mechanism to explore optimal detection models. Experimental results on three real-world datasets demonstrate the superiority of our approach.
With the rapid proliferation of 2D and 3D data, driven by advances in virtual environments and AI-generated content, cross-modal 2D-3D retrieval has attracted growing attention. However, it is easy to introduce noisy labels due to the spatial complexity of 3D content. Although various methods have been proposed to address this issue, they still struggle to handle or effectively re-exploit noisy samples. Moreover, existing approaches are prone to error accumulation due to the self-reinforcement of the model during training. To address these issues, we propose a Noise-Robust Cross-modal Learning (NRCL) framework based on the hybrid strategy. Specifically, NRCL introduces a Robust Cross-modal Co-separator (RCC), which separates noisy samples from clean ones by leveraging modality complementarity and adopting a co-teaching paradigm to mitigate potential error accumulation of the single model during training. Besides, a Reliable Soft Rectification (RSR) method is adopted to correct noisy labels by aggregating historical and dual-model predictions, exploiting the discriminative information from noisy samples. Finally, a Robust Cross-modal Prototype Learning (RCPL) is proposed to improve the discriminability of inter-class and alleviate the inherent gaps across modalities in the shared common space, which jointly leverages clean and rectified labels, thereby mitigating the detrimental impact of noisy samples. Extensive experiments are conducted on three 3D multimodal datasets to verify the effectiveness of our method by comparing it with 10 state-of-the-art methods. The code is available at https://github.com/yangaonidaye123/NRCL.
Reconstructing objects and extracting high-quality surfaces play a vital role in the real world. Current 4D representations show the ability to render high-quality novel views for dynamic objects, but cannot reconstruct high-quality meshes due to their implicit or geometrically inaccurate representations. In this paper, we propose a novel representation that can reconstruct accurate meshes from sparse image input, named Dynamic 2D Gaussians (D-2DGS). We adopt 2D Gaussians for basic geometry representation and use sparse-controlled points to capture the 2D Gaussian's deformation. By extracting the object mask from the rendered high-quality image and masking the rendered depth map, we remove floaters that are prone to occur during reconstruction and can extract high-quality dynamic mesh sequences of dynamic objects. Experiments demonstrate that our D-2DGS is outstanding in reconstructing detailed and smooth high-quality meshes from sparse inputs. The code is available at https://github.com/hustvl/Dynamic-2DGS.