• Embodied Intelligence and Visual Perception (具身智能与视觉感知)
• Intelligent Perception for Medical Robotics (医疗机器人智能感知)
• Remote Sensing Image Analysis and Intelligent Observation (遥感图像分析与智能观测)
• Underwater Image Processing and Enhancement (水下图像处理与增强)
• Visual Perception for Underwater Robotics (水下机器人视觉感知)
• Multimodal Perception and Sensor Fusion (多模态感知与信息融合)
• Cross-Domain Generalization and Sim-to-Real Transfer (跨域泛化与仿真到现实迁移)
• Robust Perception and Decision-Making in Challenging Environments (复杂环境下的鲁棒感知与决策)

This session explores how embodied visual perception can move from passive recognition to active understanding. Instead of just recognizing images, an agent actively observes, interacts, and uses action feedback to build perception that serves the task. That matters for safe decisions in degraded, changing environments like medical and underwater settings. We welcome researchers and practitioners from computer vision, robotics, embodied intelligence, medical imaging, and underwater perception. The goal is to clarify core problems, evaluation paradigms, and cross-domain transfer, and to help develop embodied vision systems and benchmarks that can be deployed in practice.