Xingxi Yin

Chenglin Li, Qianglong Chen, Feng Han et al.

Long-form video understanding remains a fundamental challenge for current Video Large Language Models. Most existing models rely on static reasoning over uniformly sampled frames, which weakens temporal localization and leads to substantial information loss in long videos. Agentic tools such as temporal retrieval, spatial zoom, and temporal zoom offer a natural way to overcome these limitations by enabling adaptive exploration of key moments. However, constructing agentic video understanding data requires models that already possess strong long-form video comprehension, creating a circular dependency. We address this challenge with VideoThinker, an agentic Video Large Language Model trained entirely on synthetic tool interaction trajectories. Our key idea is to convert videos into rich captions and employ a powerful agentic language model to generate multi-step tool use sequences in caption space. These trajectories are subsequently grounded back to video by replacing captions with the corresponding frames, yielding a large-scale interleaved video and tool reasoning dataset without requiring any long-form understanding from the underlying model. Training on this synthetic agentic dataset equips VideoThinker with dynamic reasoning capabilities, adaptive temporal exploration, and multi-step tool use. Remarkably, VideoThinker significantly outperforms both caption-only language model agents and strong video model baselines across long-video benchmarks, demonstrating the effectiveness of tool augmented synthetic data and adaptive retrieval and zoom reasoning for long-form video understanding.

Xingxi Yin, Jingfeng Zhang, Yue Deng et al.

Text-to-image (T2I) diffusion models are widely used in image editing due to their powerful generative capabilities. However, achieving fine-grained control over specific object attributes, such as color and material, remains a considerable challenge. Existing methods often fail to accurately modify these attributes or compromise structural integrity and overall image consistency. To fill this gap, we introduce Structure Preservation and Attribute Amplification (SPAA), a novel training-free framework that enables precise generation of color and material attributes for the same object by intelligently manipulating self-attention maps and cross-attention values within diffusion models. Building on SPAA, we integrate multi-modal large language models (MLLMs) to automate data curation and instruction generation. Leveraging this object attribute data collection engine, we construct the Attribute Dataset, encompassing a comprehensive range of colors and materials across diverse object categories. Using this generated dataset, we propose InstructAttribute, an instruction-tuned model that enables fine-grained and object-level attribute editing through natural language prompts. This capability holds significant practical implications for diverse fields, from accelerating product design and e-commerce visualization to enhancing virtual try-on experiences. Extensive experiments demonstrate that InstructAttribute outperforms existing instruction-based baselines, achieving a superior balance between attribute modification accuracy and structural preservation.

Xingxi Yin, Zhi Li, Jingfeng Zhang et al.

Text-to-image (T2I) diffusion models, with their impressive generative capabilities, have been adopted for image editing tasks, demonstrating remarkable efficacy. However, due to attention leakage and collision between the cross-attention map of the object and the new color attribute from the text prompt, text-guided image editing methods may fail to change the color of an object, resulting in a misalignment between the resulting image and the text prompt. In this paper, we conduct an in-depth analysis on the process of text-guided image synthesizing and what semantic information different cross-attention blocks have learned. We observe that the visual representation of an object is determined in the up-block of the diffusion model in the early stage of the denoising process, and color adjustment can be achieved through value matrices alignment in the cross-attention layer. Based on our findings, we propose a straightforward, yet stable, and effective image-guided method to modify the color of an object without requiring any additional fine-tuning or training. Lastly, we present a benchmark dataset called COLORBENCH, the first benchmark to evaluate the performance of color change methods. Extensive experiments validate the effectiveness of our method in object-level color editing and surpass the performance of popular text-guided image editing approaches in both synthesized and real images.

Xingxi Yin, Yicheng Li, Gong Yan et al.

Sequential identity consistency under precise transient attribute control remains a long-standing challenge in controllable visual storytelling. Existing datasets lack sufficient fidelity and fail to disentangle stable identities from transient attributes, limiting structured control over pose, expression, and scene composition and thus constraining reliable sequential synthesis. To address this gap, we introduce \textbf{2K-Characters-10K-Stories}, a multi-modal stylized narrative dataset of \textbf{2{,}000} uniquely stylized characters appearing across \textbf{10{,}000} illustration stories. It is the first dataset that pairs large-scale unique identities with explicit, decoupled control signals for sequential identity consistency. We introduce a \textbf{Human-in-the-Loop pipeline (HiL)} that leverages expert-verified character templates and LLM-guided narrative planning to generate highly-aligned structured data. A \textbf{decoupled control} scheme separates persistent identity from transient attributes -- pose and expression -- while a \textbf{Quality-Gated loop} integrating MMLM evaluation, Auto-Prompt Tuning, and Local Image Editing enforces pixel-level consistency. Extensive experiments demonstrate that models fine-tuned on our dataset achieves performance comparable to closed-source models in generating visual narratives.