• Image- and Video-Based Multi-View Geometry, 3D/4D Reconstruction, and Dynamic Scene Modeling (基于图像与视频的多视图几何、三维/四维重建与动态场景建模)
• Neural Scene Representations, Neural Rendering, and Novel View Synthesis (神经场景表示、神经渲染与新视角合成)
• Depth Estimation, Semantic Scene Understanding, Visual Tracking, and Processing of Point Clouds, Meshes, and Volumetric Data (深度估计、语义场景理解、视觉跟踪及点云、网格与体数据处理)
• Multi-Sensor Calibration, Spatiotemporal Alignment, and Geometry-Aware Multimodal Image and Signal Fusion (多传感器标定、时空对齐与几何感知的多模态图像和信号融合)
• Computational Imaging and Physics-Guided Inverse Problems for Digital-Twin Sensing and Reconstruction (面向数字孪生感知与重建的计算成像及物理引导逆问题)
• Data Mining, Feature Learning, and Visual Analytics for Image, Video, and Multimodal Spatiotemporal Data (面向图像、视频与多模态时空数据的数据挖掘、特征学习与可视分析)
• Digital-Twin State Recognition, Visual Anomaly Detection, Multimodal Temporal Modeling, and Future-State Prediction (数字孪生状态识别、视觉异常检测、多模态时序建模与未来状态预测)
• Real-Time Scene Updating, Online State Estimation, Multimodal Data Assimilation, and Continuous Physical–Virtual Synchronization (实时场景更新、在线状态估计、多模态数据同化与虚实持续同步)
• Visual Fidelity Assessment, Image and Signal Quality Metrics, Uncertainty Estimation, and Robust Validation for Digital Twins (数字孪生的视觉保真度评价、图像与信号质量度量、不确定性估计及鲁棒验证)
• Vision and Multimodal Digital-Twin Applications in Smart Manufacturing, Intelligent Infrastructure, Healthcare, Robotics, Cultural Heritage, and Smart Environments (数字孪生在智能制造、智慧基础设施、医疗健康、机器人、文化遗产与智慧环境中的视觉及多模态应用)

This track focuses on digital-twin technologies driven by computer vision, image processing, and signal processing, with data mining applied specifically to image, video, 3D geometry, multimodal sensing, and operational time-series data. It addresses methods for the high-fidelity 3D/4D reconstruction, state perception, continuous updating, and reliable validation of physical objects, environments, and processes. Topics include multi-view geometry, dynamic scene modeling, neural scene representations and rendering, computational imaging, multimodal calibration and fusion, spatiotemporal pattern mining, change and anomaly detection, online state estimation, and continuous physical–virtual synchronization. Particular attention is given to accuracy, robustness, real-time performance, uncertainty quantification, and verifiable fidelity in complex real-world environments. The track welcomes theoretical, methodological, and system-level contributions in smart manufacturing, infrastructure, healthcare, robotics, cultural heritage, and intelligent environments.