We are now on the way of reaching out the Metaverse, which has been coined to further facilitate the digital transformation in every aspect of our physical lives. There are many technologies, e.g., VR/AR, that drive the transition from the current internet to the Meteverse. Computer vision is an important technique that rigorously supports the functioning of any intelligent systems based on visual cues. In this talk, I will first provide a comprehensive analysis for the development of the state-of-the-art computer vision technologies, such as SLAM, object detection and segmentation, digital human modeling, etc., as they play a crucial role for the VR/AR and further for Metaverse ecosystem. Then, I will share some recent research on 3D scene understanding and reconstruction, which can be helpful and potential for the development of the Metaverse system.