Social media platforms enable users to express emotions by posting text with accompanying images. In this paper, we propose the Affective Image Filter (AIF) task, which aims to reflect visually-abstract emotions from text into visually-concrete images, thereby creating emotionally compelling results. We first introduce the AIF dataset and the formulation of the AIF models. Then, we present AIF-B as an initial attempt based on a multi-modal transformer architecture. After that, we propose AIF-D as an extension of AIF-B towards deeper emotional reflection, effectively leveraging generative priors from pre-trained large-scale diffusion models. Quantitative and qualitative experiments demonstrate that AIF models achieve superior performance for both content consistency and emotional fidelity compared to state-of-the-art methods. Extensive user study experiments demonstrate that AIF models are significantly more effective at evoking specific emotions. Based on the presented results, we comprehensively discuss the value and potential of AIF models.

While recent advancements in generative models have achieved remarkable visual fidelity in video synthesis, creating coherent multi-shot narratives remains a significant challenge. To address this, keyframe-based approaches have emerged as a promising alternative to computationally intensive end-to-end methods, offering the advantages of fine-grained control and greater efficiency. However, these methods often fail to maintain cross-shot consistency and capture cinematic language. In this paper, we introduce STAGE, a SToryboard-Anchored GEneration workflow to reformulate the keyframe-based multi-shot video generation task. Instead of using sparse keyframes, we propose STEP2 to predict a structural storyboard composed of start-end frame pairs for each shot. We introduce the multi-shot memory pack to ensure long-range entity consistency, the dual-encoding strategy for intra-shot coherence, and the two-stage training scheme to learn cinematic inter-shot transition. We also contribute the large-scale ConStoryBoard dataset, including high-quality movie clips with fine-grained annotations for story progression, cinematic attributes, and human preferences. Extensive experiments demonstrate that STAGE achieves superior performance in structured narrative control and cross-shot coherence.

Automatic video colorization is inherently an ill-posed problem because each monochrome frame has multiple optional color candidates. Existing language-based video editing methods struggle to balance content preservation and color assignment, leading to incorrect colors for corresponding instances. Alternatively, language-based image colorization methods combined with post-processing algorithms still struggle to maintain temporal consistency. To address these issues, we present Language-based video Colorization for Creative and Consistent Color (L-C4) to guide the colorization process using user-provided language descriptions. Our model is built upon a pre-trained cross-modality generative model, leveraging its comprehensive language understanding and robust color representation for creative colorization results. We introduce the luminance encoder to preserve video content and propose the cross-modality pre-fusion module for effective instance-aware colorization. Additionally, we propose temporally deformable attention to prevent flickering or color shifts, and cross-clip fusion to maintain long-term color consistency. Extensive experimental results demonstrate that L-C4 outperforms relevant methods, achieving semantically accurate colors, controllable creative correspondence, and temporally robust consistency.

While language-based video colorization addresses the inherent ambiguity of color assignment, language descriptions typically focus on central objects, neglecting the crucial context of emotional tone and surrounding environment necessary for accurate film colorization. In this paper, we introduce L-VOCAL, a novel framework for language-based video colorization that leverages audio alignment to supplement context not explicitly provided by language. L-VOCAL pretrains an alignment model to establish correspondences between color and audio, enabling the learning of emotional tone and environmental atmosphere. Subsequently, these aligned audio features guide the colorization process through specially designed condition injection modules. We additionally contribute L-VACOLOR, a new dataset tailored for this task, consisting of cinematic clips with diverse color and audio tones for training and evaluation. Extensive experimental results demonstrate that L-VOCAL produces colorization results that more accurately reflect filmmakers’ artistic expression.

Effective poster design requires rapidly capturing attention and clearly conveying messages. Inspired by the “contrast effects” principle, we propose ReContraster, the first training-free model to leverage regional contrast to make posters stand out. By emulating the cognitive behaviors of a poster designer, ReContraster introduces the compositional multi-agent system to identify elements, organize layout, and evaluate generated poster candidates. To further ensure harmonious transitions across region boundaries, ReContraster integrates the hybrid denoising strategy during the diffusion process. We additionally contribute a new benchmark dataset for comprehensive evaluation. Seven quantitative metrics and four user studies confirm its superiority over relevant state-of-the-art methods, producing visually striking and aesthetically appealing posters.

The surging demand for adapting long-form cinematic content into short videos has motivated the need for versatile automatic video compilation systems. However, existing compilation methods are limited to predefined tasks, and the community lacks a comprehensive benchmark to evaluate the cinematic compilation. To address this, we introduce CineBench, the first benchmark for instruction-driven cinematic video compilation, featuring diverse user instructions and high-quality ground-truth compilations annotated by professional editors. To overcome contextual collapse and temporal fragmentation, we present CineAgents, a multi-agent system that reformulates cinematic video compilation into “design-and-compose” paradigm. CineAgents performs script reverse-engineering to construct a hierarchical narrative memory to provide multi-level context and employs an iterative narrative planning process that refines a creative blueprint into a final compiled script. Extensive experiments demonstrate that CineAgents significantly outperforms existing methods, generating compilations with superior narrative coherence and logical coherence.

Accurate reconstruction of reflective surfaces remains a fundamental challenge in computer vision, with broad applications in real-time virtual reality and digital content creation. Although 3D Gaussian Splatting (3DGS) enables efficient novel-view rendering with explicit representations, its performance on reflective surfaces still lags behind implicit neural methods, especially in recovering fine geometry and surface normals. To address this gap, we propose PolGS++, a physically-guided polarimetric Gaussian Splatting framework for fast reflective surface reconstruction. Specifically, we integrate a polarized BRDF (pBRDF) model into 3DGS to explicitly decouple diffuse and specular components, providing physically grounded reflectance modeling and stronger geometric cues for reflective surface recovery. Furthermore, we introduce a depth-guided visibility mask acquisition mechanism that enables angle-of-polarization (AoP)-based tangent-space consistency constraints in Gaussian Splatting without costly ray-tracing intersections. This physically guided design improves reconstruction quality and efficiency, requiring only about 10 minutes of training. Extensive experiments on both synthetic and real-world datasets validate the effectiveness of our method.

Diffusion models have achieved remarkable progress in video generation, but their controllability remains a major limitation. Key scene factors such as layout, lighting, and camera trajectory are often entangled or only weakly modeled, restricting their applicability in domains like filmmaking and virtual production where explicit scene control is essential. We present LiVER, a diffusion-based framework for scene-controllable video generation. To achieve this, we introduce a novel framework that conditions video synthesis on explicit 3D scene properties, supported by a new large-scale dataset with dense annotations of object layout, lighting, and camera parameters. Our method disentangles these properties by rendering control signals from a unified 3D representation. We propose a lightweight conditioning module and a progressive training strategy to integrate these signals into a foundational video diffusion model, ensuring stable convergence and high fidelity. Our framework enables a wide range of applications, including image-to-video and video-to-video synthesis where the underlying 3D scene is fully editable. To further enhance usability, we develop a scene agent that automatically translates high-level user instructions into the required 3D control signals. Experiments show that LiVER achieves state-of-the-art photorealism and temporal consistency while enabling precise, disentangled control over scene factors, setting a new standard for controllable video generation.

Music-driven dance video synthesis aims to animate a reference person according to a given music clip. The task is challenging because it requires a model to jointly learn music-to-motion correspondence, identity-preserving human animation, temporal coherence, and visually realistic video generation. We present FlowDance, a music-driven dance video generation framework that integrates explicit motion modeling with reference-preserving visual synthesis through parallel pose and RGB streams. We further introduce timestep-aware pose injection to adapt structural guidance across denoising steps and persistent identity injection to preserve the reference appearance over long video. To support this task, we further build a popularity-curated, high-resolution in-the-wild dance video dataset with synchronized music, RGB videos, 3D body motion, camera parameters, and projected 2D pose annotations. Extensive experiments show that FlowDance achieves strong performance in both dance motion generation and music-driven dance video synthesis.

Language-guided image colorization has achieved remarkable progress by leveraging multimodal priors and diffusion models. However, developing a unified model capable of handling diverse image domains remains a formidable challenge. The primary obstacle lies in the significant discrepancy between spatially dense inputs (e.g., natural photography), which demand strict structural fidelity, and spatially sparse inputs (e.g., anime sketches), which require strong generative capabilities to populate semantic colors. Existing “one-size-fits-all” models or standard Mixture of Experts (MoE) architectures with homogeneous experts often struggle to reconcile these divergent needs, leading to domain color drift or structural degradation. To address this, we propose CROME (Cross-domain Optimal Mixture of Experts), a novel framework designed for universal colorization. Unlike previous approaches, CROME introduces heterogeneous experts differentiated by granularity and modeling capacity. We design global-level experts for domain style adaptation and patch-level experts for fine-grained semantic refinement. Furthermore, the model dynamically allocates computational resources based on information density: high-capacity experts are routed to handle generative tasks for sparse inputs, while low-capacity experts preserve fidelity for dense inputs. Facilitated by a domain-aware routing mechanism, CROME successfully achieves high-quality, controllable colorization across six diverse domains within a single unified model.

Audio is inherently temporal and closely synchronized with the visual world, making it a naturally aligned and expressive control signal for controllable video generation (e.g., movies). Beyond control, directly translating audio into video is essential for understanding and visualizing rich audio narratives (e.g., Podcasts or historical recordings). However, existing approaches fall short in generating high-quality videos with precise audio-visual synchronization, especially across diverse and complex audio types. In this work, we introduce MTV, a versatile framework for audio-sync video generation. MTV explicitly separates audios into speech, effects, and music tracks, enabling disentangled control over lip motion, event timing, and visual mood, respectively—resulting in fine-grained and semantically aligned video generation. To support the framework, we additionally present DEMIX, a dataset comprising high-quality cinematic videos and demixed audio tracks. DEMIX is structured into five overlapped subsets, enabling scalable multi-stage training for diverse generation scenarios. Extensive experiments demonstrate that MTV achieves state-of-the-art performance across six standard metrics spanning video quality, text-video consistency, and audio-video alignment.

Video Temporal Grounding (VTG) involves locating the start and end times of a video clip based on a given textual query. Existing methods face challenges including temporal boundary localization bias and class imbalance in points classification. To alleviate these, we propose a Dual Fusion Network (DF-Net) in which cross-modal fusion processes occur in the joint encoder and decoder. In the encoder, we design a multi-step sampling strategy and knowledge aggregation module to extract and aggregate features from central frames and neighbors and undergo the first cross-model fusion. In decoding, we employ a transformer decoder to design the second cross-modal fusion and propose a novel Kullback-Leibler (KL) divergence-based objective function combined with focal and distance-based regression losses. Experiments on short and long video benchmarks demonstrate significant performance gains.

In daily life, images as common affective stimuli have widespread applications. Despite significant progress in text-driven image editing, there is limited work focusing on understanding users’ emotional requests. In this paper, we introduce AIEdiT for Affective Image Editing using Text descriptions, which evokes specific emotions by adaptively shaping multiple emotional factors across the entire image. To represent universal emotional priors, we build the continuous emotional spectrum and extract nuanced emotional requests. To manipulate emotional factors, we design the emotional mapper to translate visually-abstract emotional requests to visually-concrete semantic representations. To ensure that editing results evoke specific emotions, we introduce an MLLM to supervise the model training. During inference, we strategically distort visual elements and subsequently shape corresponding emotional factors to edit images according to users’ instructions. Additionally, we introduce a large-scale dataset that includes the emotion-aligned text and image pair set for training and evaluation. Extensive experiments demonstrate that AIEdiT achieves superior performance, effectively reflecting users’ emotional requests.

With the rapid development of Large Vision Language Models, the focus of Graphical User Interface (GUI) agent tasks shifts from single-screen tasks to complex screen navigation challenges. However, real-world GUI environments, such as PC software and mobile Apps, are often complex and proprietary, making it difficult to obtain the comprehensive environment information needed for agent training and evaluation. This limitation hinders systematic investigation and benchmarking of agent navigation capabilities. To address this limitation, we introduce GUI Exploration Lab, a simulation environment engine for GUI agent navigation research that enables flexible definition and composition of screens, icons, and navigation graphs, while providing full access to environment information for comprehensive agent training and evaluation. Through extensive experiments, we find that supervised fine-tuning enables effective memorization of fundamental knowledge, serving as a crucial foundation for subsequent training. Building on this, single-turn reinforcement learning further enhances generalization to unseen scenarios. Finally, multi-turn reinforcement learning encourages the development of exploration strategies through interactive trial and error, leading to further improvements in screen navigation performance. We validate our methods on both static and interactive benchmarks, demonstrating that our findings generalize effectively to real-world scenarios. These findings demonstrate the advantages of reinforcement learning approaches in GUI navigation and offer practical guidance for building more capable and generalizable GUI agents.

In this paper, we introduce OpenCIR, a fullyfunctional Conditional Image Repainting (CIR) model designed for local image editing. Given an image and a combination of conditions related to geometry, texture, and color, CIR models are required to repaint instances and seamlessly composite them with the original images. Previous CIR models suffer from limited object categories, restricted condition modalities, and demanded geometry precision. In contrast, leveraging the generative priors from pre-trained models, OpenCIR could repaint open object categories. Equipped with redesigned condition injection modules and the condition extension strategy, OpenCIR is able to understand open condition modalities. Adopting the contour refinement strategy, OpenCIR allows users to specify instances with open geometry precision. In addition, we contribute the OPEN-CIR dataset, which includes detailed annotations, tailored for the comprehensive training and evaluation of the OpenCIR model. Extensive experiments demonstrate that OpenCIR outperforms relevant state-of-the-art methods, achieving superior visual quality, and more favorable results by human evaluators.

Event Extraction (EE) in low-resourced environments aims to derive structured representations of event information from limited annotations. Recent approaches have focused on ontology-agnostic parameterization to address this challenge, treating specific event ontology information as textual input instead of model parameters. However, these methods highly rely on semantic understanding ability to support cross-type transferability, while their fixed ontology modeling is essentially exclusive across seen and unseen types, their cross-type transferability would suffer during fine-tuning. Motivated by ideas of imitation learning in the educational field, we propose a Dynamic Demonstration-based Learning Framework for Event Extraction(D2E2), achieving event extraction in a rank-then-imitate manners. As imitation ability relies less on semantic understanding ability and can be jointly learned across types, it is supposed that D2E2 utilize limited annotations more efficiently. During training of D2E2, similar annotated examples are sampled based on BM25 algorithm, composing a series of positive or negative pairs. Then these pairs are used to fine-tune a semantic-based retriever, and demonstration-based learning extractor to extract events via imitation. At inference time, the semantic-based retriever first retrieves annotated examples similar with candidate events in the input text, then the extractor uses them as demonstrations to conceptualize and structure triggers and arguments in the input text via imitating those in annotations. Experimental evaluations conducted on public benchmarks ACE05 and FewFC demonstrate the effectiveness of our method. Results indicate significant performance improvements over existing approaches in low-resourced scenarios, while also achieving competitive or superior performance in high-resourced scenarios.

Polarization images facilitate image enhancement and 3D reconstruction tasks, but the limited accessibility of polarization cameras hinders their broader application. This gap drives the need for synthesizing photorealistic polarization images.The existing polarization simulator Mitsuba relies on a parametric polarization image formation model and requires extensive 3D assets covering shape and PBR materials, preventing it from generating large-scale photorealistic images. To address this problem, we propose PolarAnything, capable of synthesizing polarization images from a single RGB input with both photorealism and physical accuracy, eliminating the dependency on 3D asset collections. Drawing inspiration from the zero-shot performance of pretrained diffusion models, we introduce a diffusion-based generative framework with an effective representation strategy that preserves the fidelity of polarization properties. Experiments show that our model generates high-quality polarization images and supports downstream tasks like shape from polarization.

We introduce VIRES, a video instance repainting method with sketch and text guidance, enabling video instance repainting, replacement, generation, and removal. Existing approaches struggle with temporal consistency and accurate alignment with the provided sketch sequence. VIRES leverages the generative priors of text-to-video models to maintain temporal consistency and produce visually pleasing results. We propose the Sequential ControlNet with the standardized self-scaling, which effectively extracts structure layouts and adaptively captures high-contrast sketch details. We further augment the diffusion transformer backbone with the sketch attention to interpret and inject fine-grained sketch semantics. A sketch-aware encoder ensures that repainted results are aligned with the provided sketch sequence. Additionally, we contribute the VIRESET, a dataset with detailed annotations tailored for training and evaluating video instance editing methods. Experimental results demonstrate the effectiveness of VIRES, which outperforms state-of-the-art methods in visual quality, temporal consistency, condition alignment, and human ratings. The code, dataset and pretrained models are available at: https://hjzheng.net/projects/VIRES.

This report describes our model submitted for the EvaHan 2025 shared task on named entity recognition for ancient Chinese literary works. Since we participated in the task of closed modality, our method is based on the appointed pretrained language model GujiRoBERTajian-fan and we used appointed datasets. We carried out experiments on decoding strategies and schedulers to verify the effect of our method. In the final test, our method outperformed the official baseline, demonstrating its effectiveness. In the end, for the results, this report gives an analysis from the perspective of data composition.

Photographic style, as a composition of certain photographic concepts, is the charm behind renowned photographers. But learning and transferring photographic style need a profound understanding of how the photo is edited from the unknown original appearance. Previous works either fail to learn meaningful photographic concepts from reference images, or cannot preserve the content of the content image. To tackle these issues, we proposed a Personalized Image Filter (PIF). Based on a pretrained text-to-image diffusion model, the generative prior enables PIF to learn the average appearance of photographic concepts, as well as how to adjust them according to text prompts. PIF then learns the photographic style of reference images with the textual inversion technique, by optimizing the prompts for the photographic concepts. PIF shows outstanding performance in extracting and transferring various kinds of photographic style. Project page: https://pif.pages.dev/