3D video is the ultimate image medium recording dynamic visual events in the real world as is. Recorded object behaviors can be observed from any viewpoint, because 3D video records the object's full 3D shape, motion, and precise surface properties (i.e. color and texture). In this paper, we present research attainments so far obtained: 1. a PC cluster system for real-time reconstruction of dynamic 3D object behaviors from multi-view video images, 2. an active 3D mesh model for reconstructing the accurate 3D object shape, 3. an algorithm of generating natural-looking texture on the 3D object surface from the multi-view video images, and 4. an editing system for visualizing 3D video with an omni-directional background image using versatile 3D camera works. Experimental results demonstrate the effectiveness of these methods in generating high fidelity and attractive object images from arbitrary viewpoints.